CA2313812C - Process for correcting atmospheric influences in multispectral optical remote sensing data - Google Patents
Process for correcting atmospheric influences in multispectral optical remote sensing data Download PDFInfo
- Publication number
- CA2313812C CA2313812C CA002313812A CA2313812A CA2313812C CA 2313812 C CA2313812 C CA 2313812C CA 002313812 A CA002313812 A CA 002313812A CA 2313812 A CA2313812 A CA 2313812A CA 2313812 C CA2313812 C CA 2313812C
- Authority
- CA
- Canada
- Prior art keywords
- module
- data
- sub
- analysis
- process according
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 114
- 230000003287 optical effect Effects 0.000 title claims abstract description 9
- 238000012937 correction Methods 0.000 claims abstract description 46
- 238000004364 calculation method Methods 0.000 claims abstract description 10
- 238000006243 chemical reaction Methods 0.000 claims abstract description 5
- 230000003595 spectral effect Effects 0.000 claims description 30
- 230000036961 partial effect Effects 0.000 claims description 16
- 238000001228 spectrum Methods 0.000 claims description 13
- 230000007717 exclusion Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 8
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 8
- 239000000443 aerosol Substances 0.000 claims description 7
- 230000003247 decreasing effect Effects 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 4
- 230000005855 radiation Effects 0.000 claims description 4
- 238000012546 transfer Methods 0.000 claims description 4
- 101150082208 DIABLO gene Proteins 0.000 claims description 3
- 102100033189 Diablo IAP-binding mitochondrial protein Human genes 0.000 claims description 3
- CBENFWSGALASAD-UHFFFAOYSA-N Ozone Chemical compound [O-][O+]=O CBENFWSGALASAD-UHFFFAOYSA-N 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 2
- 230000007704 transition Effects 0.000 claims description 2
- 238000002329 infrared spectrum Methods 0.000 claims 2
- 230000004069 differentiation Effects 0.000 claims 1
- 238000010348 incorporation Methods 0.000 abstract description 2
- 101000928995 Caenorhabditis elegans Putative deoxyribose-phosphate aldolase Proteins 0.000 abstract 2
- 102100037802 Deoxyribose-phosphate aldolase Human genes 0.000 abstract 2
- 230000002452 interceptive effect Effects 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 239000006101 laboratory sample Substances 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 230000000873 masking effect Effects 0.000 description 1
- 239000011435 rock Substances 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000000153 supplemental effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C11/00—Photogrammetry or videogrammetry, e.g. stereogrammetry; Photographic surveying
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration by the use of more than one image, e.g. averaging, subtraction
-
- G06T5/73—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
- G06T2207/10036—Multispectral image; Hyperspectral image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
Abstract
The inventive process for correcting atmospheric influences in multispectral optical remote sensing data, which are acquired as raw data in satellite or airborne sensors for earth observation, comprises the combination of a pre-classification (DERA) of the raw data for an automatic identification of predefined classes, a correction calculation (CORA) for a conversion of the uncorrected to corrected reflectances on the ground, including an incorporation of current atmospheric data for a precise description of the atmospheric condition. The pre-classification (DERA) permits a more precise correction calculation (CORA) by generating required a priori knowledge. The method has applications in satellite or airborne remote sensing of the earth's surface.
Description
PROCESS FOR CORRECTING ATMOSPHERIC INFLUENCES
IN MUI~TISPE~AL OPTICAL REMOTE SENSING DATA
BACKGROUND QF THE INVENTION
Technical Field of the Invention The present invention relates to a process for correcting atmospheric influences in multispectral optical remote sensing data that are acquired in different types of satellite or airborne sensors for earth observation with different geometric and/or spectral resolutions, and read in and precessed as raw data to generate an image.
D v s .-. v. T ,.. i-A fundamental prerequisite for deriving quantitative parameters and indicators from remotely sensed data, apart from geo-referencing, is the atmospheric correction.
A number of basic processes for the atmospheric correction of multispectral remote sensing data already exists. However, some of these processes are only on a scientific level of development (laboratory samples, e.g., H.
Rahman, G. Dedieu: "SMAC '; A simplified method for atmospheric correction of satellite measurements in the solar spectrum, Int. J. Rem. Sens., 15, l, pages 123-143, 1994, and INEXACT ", Th. Popp: Correcting atmospheric masking to retrieve the spectral albedo of land surfaces from satellite measurements, Int. J. Rem. Sens., 16, pages 3483-3508, 1995), which does not permit a routine automatic pro~Gessing of multispectral remote sensing data from a great variety of sensors.
Then there are processes that are currently being used as industrial samples in commercial program packages for the atmospheric correction of multispectral remote sensing data, one of which is known from DE 41 02 579 C2. They are based on the determination of reference areas of a low reflectance, which must be identified in the remotely sensed data. The criteria that are used for these reference areas may be the grey value, the color, or the multispectral signature. The calculation of the atmospheric correction furthermore requires that the properties of the atmosphere in need of correction be known. As a rule, this is done by entering pre-set standard atmospheres. These processes, which are being used as industrial samples, require the interactive interaction, for example for the selection of reference areas and atmospheric parameters, by an expert who must possess specialized knowledge and experience in the field of atmospheric correction. These processes, therefore, cannot be used for an automatic atmospheric correction of remotely sensed data.
IN MUI~TISPE~AL OPTICAL REMOTE SENSING DATA
BACKGROUND QF THE INVENTION
Technical Field of the Invention The present invention relates to a process for correcting atmospheric influences in multispectral optical remote sensing data that are acquired in different types of satellite or airborne sensors for earth observation with different geometric and/or spectral resolutions, and read in and precessed as raw data to generate an image.
D v s .-. v. T ,.. i-A fundamental prerequisite for deriving quantitative parameters and indicators from remotely sensed data, apart from geo-referencing, is the atmospheric correction.
A number of basic processes for the atmospheric correction of multispectral remote sensing data already exists. However, some of these processes are only on a scientific level of development (laboratory samples, e.g., H.
Rahman, G. Dedieu: "SMAC '; A simplified method for atmospheric correction of satellite measurements in the solar spectrum, Int. J. Rem. Sens., 15, l, pages 123-143, 1994, and INEXACT ", Th. Popp: Correcting atmospheric masking to retrieve the spectral albedo of land surfaces from satellite measurements, Int. J. Rem. Sens., 16, pages 3483-3508, 1995), which does not permit a routine automatic pro~Gessing of multispectral remote sensing data from a great variety of sensors.
Then there are processes that are currently being used as industrial samples in commercial program packages for the atmospheric correction of multispectral remote sensing data, one of which is known from DE 41 02 579 C2. They are based on the determination of reference areas of a low reflectance, which must be identified in the remotely sensed data. The criteria that are used for these reference areas may be the grey value, the color, or the multispectral signature. The calculation of the atmospheric correction furthermore requires that the properties of the atmosphere in need of correction be known. As a rule, this is done by entering pre-set standard atmospheres. These processes, which are being used as industrial samples, require the interactive interaction, for example for the selection of reference areas and atmospheric parameters, by an expert who must possess specialized knowledge and experience in the field of atmospheric correction. These processes, therefore, cannot be used for an automatic atmospheric correction of remotely sensed data.
In the commercially applied industrial sample processes, the atmospheric correction is performed manually through interactive parameter adjustments and, as a rule, this is done using predefined standard information, e.g., in the form of a limited number of standard atmospheres and/or a predefined visibility. Selecting the best-suited standard information for the given remotely sensed data set being processed requires expert knowledge on the part of the operating personnel. Furthermore, until now there is no automatic identification of validation areas, e.g., of reference areas of a low reflectance and of areas of known reflectance behavior. These areas are currently also identified and marked interactively by the operating personnel.
The aforementioned scientific processes in the form of known laboratory samples are generally optimized with respect to a specific sensor or even to a specific application, or they utilize only supplemental data that are poorly correlated with respect to time/space, e.g., climatologies and weather analysis data.
Furthermore, in the known industrial samples and most laboratory samples, the anisotropy of the reflectance on the ground is not taken into consideration. A further, hitherto unsolved problem in the processing of remotely sensed ', data lies in the fact that, while it is true that measurements of the current atmospheric condition can be incorporated for the correction of individual data~'sets, as a rule, a large-scale incorporation of current atmospheric parameters can not take place within these processes.
The influence of the non-inclusion of the atmospheric parameters can be demonstrated, for example, for the. so-called normalized differential vegetation index (NDVI).
The NDVI is obtained from bi-spectral measurements in the red (channel 1) and in the near-infrared (channel 2) and represents a standard value which, because of the method by which it is calculated, already provides a correction of the zeroth order of the atmospheric influence. The following table provides an overview of the possible influence of the most important atmospheric parameters (ozone, water vapor, molecule or Rayleigh scattering, aerosol scattering) on data of the spectral reflectance and the NDVI based on the example of a known sensor (NOAA-AVHRR) and thus demonstrates the errors that can still be attached to this correction of the zeroth order if current atmospheric parameters are not used for the atmospheric correction. The proportional effects (transmission) are listed in the table in percentages and other information in absolute reflectances.
The aforementioned scientific processes in the form of known laboratory samples are generally optimized with respect to a specific sensor or even to a specific application, or they utilize only supplemental data that are poorly correlated with respect to time/space, e.g., climatologies and weather analysis data.
Furthermore, in the known industrial samples and most laboratory samples, the anisotropy of the reflectance on the ground is not taken into consideration. A further, hitherto unsolved problem in the processing of remotely sensed ', data lies in the fact that, while it is true that measurements of the current atmospheric condition can be incorporated for the correction of individual data~'sets, as a rule, a large-scale incorporation of current atmospheric parameters can not take place within these processes.
The influence of the non-inclusion of the atmospheric parameters can be demonstrated, for example, for the. so-called normalized differential vegetation index (NDVI).
The NDVI is obtained from bi-spectral measurements in the red (channel 1) and in the near-infrared (channel 2) and represents a standard value which, because of the method by which it is calculated, already provides a correction of the zeroth order of the atmospheric influence. The following table provides an overview of the possible influence of the most important atmospheric parameters (ozone, water vapor, molecule or Rayleigh scattering, aerosol scattering) on data of the spectral reflectance and the NDVI based on the example of a known sensor (NOAA-AVHRR) and thus demonstrates the errors that can still be attached to this correction of the zeroth order if current atmospheric parameters are not used for the atmospheric correction. The proportional effects (transmission) are listed in the table in percentages and other information in absolute reflectances.
'.: ', , Table Ozone Water Vapor Rayleigh Aerosol 250-500 0.5-9.0 1013.25 hPa Continental [D.U. ] [g/cm2J issonm 0.05-0.8 Channel 1 - - + +
620 120 nm 4-13.5$ 0.7-4.9$ 0.018-0.07 0.005-0.12 Channel 2 - - + +
885 195 nm 0.02-0.5$ 7.7-22$ 0.006-0.04 0.003-0.083 NDVI 0.05 + - - -(bare ground) 0.02-0.07 0.011-0.12 0.036-0.094 0.006-0.085 pl=0.19/p2=0.21 NDVI 0.85 + - - -(deciduous forest)0.006-0.0170.036-0.038 0.086-0.26 0.022-0.39 pl=0.03/p2=0.36 OBJECT AND SUMMARY OF THE INVENTION
The present invention is based on the aim of creating a process for correcting atmospheric influences for multispectral remote sensing data that is suitable for integration into an automatic processing chain and, in contrast to processes of the prior art, therefore meets important criteria in such a way that current reference areas are determined automatically and current atmospheric parameters are used, that no interactive involvement of the operating personnel must be required, and that no expert knowledge should be required on the part of the operating personnel.
620 120 nm 4-13.5$ 0.7-4.9$ 0.018-0.07 0.005-0.12 Channel 2 - - + +
885 195 nm 0.02-0.5$ 7.7-22$ 0.006-0.04 0.003-0.083 NDVI 0.05 + - - -(bare ground) 0.02-0.07 0.011-0.12 0.036-0.094 0.006-0.085 pl=0.19/p2=0.21 NDVI 0.85 + - - -(deciduous forest)0.006-0.0170.036-0.038 0.086-0.26 0.022-0.39 pl=0.03/p2=0.36 OBJECT AND SUMMARY OF THE INVENTION
The present invention is based on the aim of creating a process for correcting atmospheric influences for multispectral remote sensing data that is suitable for integration into an automatic processing chain and, in contrast to processes of the prior art, therefore meets important criteria in such a way that current reference areas are determined automatically and current atmospheric parameters are used, that no interactive involvement of the operating personnel must be required, and that no expert knowledge should be required on the part of the operating personnel.
In accordance with the invention, there is provided a process for correcting atmospheric influences in multispectral optical remote sensing data that are acquired by different satellite or airborne sensors for earth observations with different geometric and/or spectral resolutions, and read in and processed as raw data to generate an image, comprising the steps of:
pre-classification of the raw data for an automatic recognition of pre-defined classes performing in a first partial process; and performing a correction calculation to convert the uncorrected reflectances into corrected reflectances on the ground in a second partial process; and incorporating into the process current and essentially complete supplementary data on the current atmospheric conditions; and accessing a database (DABA) with methods and parameters in said first partial process and said second partial process, wherein in said database data are available regarding sensor parameters, model spectra, aerosol models, anisotropy types, atmosphere models, together with methods regarding radiative transfer methods, parameterization schemes, assimilation and interpolation methods;
wherein said first partial process is carried out by a first main module (DERA) detecting and identifying dark areas and areas of significant spectral behavior in raw remote sensing data, and said second partial process is carried out by a second main module (CORA) correcting calculations;
said first main module (DERA) consisting of two sub-modules (SSA, GSA), a first one (SSA) of which is responsible for a spectral signature analysis and is used for the identification of reference areas of low reflectance, e.g. of water surfaces and dark forest areas, as well as for the identification of exclusion areas, e.g. clouds and cloud shadows, and a second one (GSA) for a geometric structure analysis, wherein data from different sensors with different geometric and/or spectral resolutions are read as raw data into said first sub-module (SSA), which is aimed at an analysis of the spectral behavior on the pixel level, and processed, and an assignment of the pixels of read-in remotely sensed data set to probably detected remotely sensable objects in reality is made on the basis of generalized model spectra for remotely sensable objects on a pixel level, in such a way that model spectra required for this analysis are stored in the database (DABA), which is accessed during the first partial process, and that in said second sub-module (GSA) of said first main module (DERA) a homogeneity analysis of identified reference areas, an area size analysis, as well as an analysis of the direct and indirect neighborhood is performed in such a way that data areas that were detected by said first sub-module (SSA) are routed to said second sub-module (GSA) and examined for sufficient size; and said second main module (CORA) consisting of two sub-modules (INPRE, DACO), wherein a processing of the required supplementary data is performed in said first sub-module (INPRE) in such a manner that internal supplementary data as they are derived from the raw data and results from said first main module (DERA), are accessed in a first preparation unit (INPRE-INT), and external supplementary data, which are provided via an external interface (EXT-INTF), are processed in a second preparation unit (TNPRE-EXT) in such a manner that standard data assimilation and interpolation methods are used, which are made available in the database (DABA) as methods, and that the actual correction steps are performed in said - 6a -second sub-module (DACO) of said second main module (CORA) with the aid of the supplementary data from said first sub-module (INPRE) in such a way that correction processes are accessed in said method-containing database (DABA).
Pre-tabulated/parameterized radiative transfer calculations make the inventive process fast and, therefore, suitable for operational applications. To attain a good time-space correlation of the atmospheric data with the data in need of correction, these values are estimated, as far as possible, from the data in need of correction. Additional supplementary data that cannot be obtained from the data in need of correction can be acquired externally from operational processing chains via an external interface, and interpolated with suitable methods.
In the numerical process for an automatic atmosphere correction according to the invention, the data from different - 6b -sensors with different geometric and/or spectral resolution may be read in and processed as raw data, e.g, NOAA-AVHRR, ERS-ATSR, (SEA)WIFS, EOS-MODIS, Landsat-TM and Landsat-MSS, IRS-LISS, SPOT-HRV. Essential in the inventive process is the combination of an event-controlled classification and object identification, i.e., a localization and content-based correlation of objects, the actual correction calculation, and the use of current and complete supplementary data regarding the atmospheric condition. Only with this combination can an automatic atmospheric correction take place without interactive intervention or expert knowledge.
The inventive process is, therefore, composed of two partial processes, which are carried out by main modules. The first main. module is used for the detection and identification of dark areas and areas of significant spectral behavior in the remotely sensed data, and the second main module is used for the atmospheric correction of the remotely sensed data.
The two main modules consist of sub-modules. These are supplemented by a database in which basic static and dynamic data, as well as a priori knowledge, e.g., spectral signatures, sensor specifications, statistical properties, correction methods and assimilation methods are stored. This database is accessible by both main modules.
The first main module for its part consists of two sub-modules. The first of these two sub-modules is used for the identification of reference areas of low reflectance, e.g., of water surfaces and dark forest areas, as well as for the identification of exclusion areas, e.g., clouds and cloud shadows. In the process, this sub-module uses the model spectra and sensor-specific information stored in the database. The second one of these two sub-modules is used to perform the homogeneity analysis for identified reference areas (test for representativeness-of the selected areas) and the area size analysis, on one hand, and the analysis of the direct and indirect neighbourhood on the other hand.
The second main module, which is thus used for the atmospheric correction of the remotely sensed data, for its part also consists of two sub-modules. The first one of these two sub-modules is used for processing the required supplementary data. In the process, this sub-module accesses internal supplementary data, as they are derived from the raw data and the results from the first main module, and processes external supplementary data, which are made available via an external interface, e.g., online or via CD-ROM. This may be done using standard data assimilation and interpolation methods, which are made available in the methods database.
_ g _ The second one of these two sub-modules performs the actual correction steps with the aid of the supplementary data-from the first sub-module. In the process, correction methods may be accessed that are stored in the methods database.
BRIEF DESCRIPTION OF THE DRAWINGS
The inventive process for correcting atmospheric influences in multispectral optical remote sensing data will be explained in detail below, based on figure 1 which gives a schematic overview.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS) OF THE
TAT~1G'ATTTIITM
The process shown in the figure for an automatic atmospheric correction according to the invention consists of two different partial processes, which are combined as main modules DERA and CORA. The main module DERA ("Detection of Reference Areas") is used for the detection and identification of dark areas and areas of significant spectral behavior in the remotely sensed data, and the main module CORA
("Correction of Atmosphere") is used for the atmospheric correction of the remotely sensed data. Both main modules DERA and CORA access a database DABA with stored methods and parameters. The process provides the result in the form of a pixel image of the atmosphere-corrected ground reflectance _ g _ with incorporated supplementary atmospheric data,anisotropic reflectance characteristics of different ground type classes, and terrain elevation. Data from different sensors with different geometric and/or spectral resolutions can be read in as raw data and processed.
The main module DERA consists of two sub-modules SSA
("Spectral Signature Analysis") and GSA ("Geometric Structure Analysis"). The spectral signature analysis in the sub-module SSA is aimed at an analysis of the spectral behavior of the multispectral data set on the pixel level. The data from different sensors and/or different geometric resolutions can be read in and processed as raw data. In the process, the pixels of the remotely sensed data set are assigned to the probably detected remotely sensable objects in reality on the basis of generalized model spectra for remotely sensable objects on the pixel level. The model spectra that are required for this analysis are stored in the database DABA, which is used by the algorithm to be executed in the sub-module SSA. Also stored in the database are the sensor specifications that are required~for this analysis, which are made available as a priori knowledge.
On the basis of this information and the spectral behavior of the individual pixels, they are assigned to suitable reference areas, which may serve as dark areas. They furthermore form the basis for the assignment of a ground type class (forest, grassland, bog, savannah, uncultivated field, bushland, water, residential settlement, rock, sand , based on which a suitable function for modeling the anisotropic ground reflectance can be selected in the main module CORA.
The sub-module GSA is used for the homogeneity analysis of identified reference areas, for the area size analysis, as well as for the analysis of the direct and indirect neighbourhood. The data areas detected by the sub-module SSA used for the spectral signature analysis are routed to the sub-module GSA used for the geometric structure analysis, and examined for sufficient size. The geometric structure analysis that is performed in the sub-module GSA, in turn, is divided into a module HMA used for the homogeneity analysis, a module GFA used for the size and shape analysis of identified data areas, and a module NBA used for the neighbourhood analysis.
The module HMA for the homogeneity analysis is used for the identification of contiguous areas, or areas that must be treated separately during the further image processing.
'this is done by examining image areas for their local homogeneity with a filter matrix (m ~ m) and comparing them to threshold values. The measure for the homogeneity, which is then made available as the threshold value for the data set, is derived directly from the data. For this purpose the data set is divided into data sectors and subjected to a statistical analysis with filters of decreasing sizes. The mean value, the standard deviation, as well as the variation coefficient are derived as statistical measures.
In the next module GFA used for the size and shape analysis of the identified data areas, an object identification is assigned to these areas, so that the shape parameters can be directly assigned to the objects. For this purpose, the size of the area and the circumference and compactness of the objects are determined, so that a criterion is derived for the evaluation of the objects as reference areas.
Decision criteria for reference areas are defined as follow-s: reflectance value below a maximum size in the mid-infrared or, alternatively, near-infrared value below a spectral threshold in combination with exceeding a minimum value for a vegetation index, and, additionally, exceeding a minimum area size.
The direct and indirect neighbourhood of the data regions identified as homogenous is subsequently analyzed by the module NBA used for the neighbourhood analysis. The goal of this analysis is to identify so-called mixed pixels and interference pixels, which must be treated separately with respect to their belonging to the adjacent objects to be able to assign to them a suitable ground type class. For this purpose the data are analyzed, with filters of decreasing sizes, with respect to the transition contrasts between the identified data sectors, in order to thus be able to estimate e.g., the influence of clouds or haze and to separate phenomena of different length scales (e. g., small-scale variability of the land use from large-scale change in the atmospheric conditions).
The main module CORA consists of two sub-modules INPRE ("Input Preparation") and DACO ("Data Correction").
The sub-module INPRE is aimed at making the required supplementary data for the sub-module DACO available for each pixel and it is divided into two preparation modules INPRE-INT
and INPRE-EXT. In the process; the module INPRE-INT uses results from the first main module DERA (internal supplementary data); the module INPRE-EXT processes the atmospheric data (external supplementary-data)that are made available via an external interface EXT-IdTF(online or from CD-ROM, e.g., from the German Remote Sensing Data Center, DFD). In the process, the sub-module INPRE determines the following supplementary data for each pixel: turbidity mask, ground mask and exclusion mask (INPRE-INT), trace gas masks and terrain model (INPRE-EXT ) .
With the module INPRE-INT, the aerosol-optical thickness in visible channels above the reference areas determined by the first main module DERA is determined by means of a dark field method and transferred, with suitable interpolation routines, from the database DABA to all pixels and spectral channels (turbidity mask). During the spatial interpolation, the contiguous areas of similar atmospheric conditions that were determined by the first main module DERA
are taken into consideration. The ground type class determined by the first main module DERA is used to select the suitable model function of the anisotropic reflectance properties (ground mask). Cloud and shadow areas determined by the first main module DERA are annotated as "pixels not to be corrected (exclusion mask).
Via the online interface EXT-INTF or via CD-ROM, satellite or airborne data of total ozone_column and water vapor column (e. g., from the German Remote Sensing Data Center, DFD) are loaded by the module INPRE-EXT and converted with suitable assimilation processes (e. g., Harmonic analysis, Kalman filter, Kriging) from the database DABA to the point in time/geographical location of the raw data (trace gas masks).
A suitable section of a digital elevation model is acquired via the same interface and re-projected to the sensor coordinates.
The sub-module DACO performs the pixel-by-pixel correction of all non-excluded pixels (exclusion mask) with the aid of the supplementary data (masks) from the sub-module INPRE. In the process access is made, as desired, to quick correction methods from the database DABA (by means of known radiation transportation programs, e.g., 5S, SOS, MODTRAN
calculated lookup tables, or published parameterization schemes, e.g., SMAC, EXACT).
The sub-module DACO first performs, for each pixel of the raw data, a conversion of the value measured at the top of the atmosphere into a reflectance value on the ground using a module RECD (Reflectance Conversion). The pixel values of the turbidity mask, the ground mask, the trace gas mask and the elevation model are used in the process as supplementary data. The sub-module DACO incorporates the anisotropy of the reflectance from the earth's surface by using a suitable model function for each of the ground types determined from the main module DERA.
For the correction of the incident radiation into the instantaneous field of view of the sensor from adjacent pixels a simple adjacency filter is applied in a module ACO ("Adjacency Correction") using the turbidity mask. This is done in a second step, which, however, is necessary only for high-resolution sensors.
The modules access the methods and parameters database DABA. In this database, data (sensor parameters, model spectra, aerosol models, 'anisotropy types, atmospheric models) are available together with methods ( radiative transfer methods, parameterization schemes, assimilation and interpolation methods).
The raw data must be roughly (approximately ~ 1 pixel or ~ degree) annotated with the geographic position (geographic longitude, geographic latitude) and the observation geometry (observation zenith, observation azimuth) of each individual pixel; only "nearest neighbor" methods should be used as interpolation methods, if need be. The raw data must be multispectral and have at least one visible and one near-infrared channel. The optimum is an additional channel in the medium infrared (more exact dark field method) and a further channel in the visible (more precise spectral interpolation of the aerosol optical thickness.) Together with the observation data the precise observation time must be made available, to be able to calculate the position of the sun.
Alternately, the zenith and azimuth angle of the sun may be provided for each pixel as an additional channel.
pre-classification of the raw data for an automatic recognition of pre-defined classes performing in a first partial process; and performing a correction calculation to convert the uncorrected reflectances into corrected reflectances on the ground in a second partial process; and incorporating into the process current and essentially complete supplementary data on the current atmospheric conditions; and accessing a database (DABA) with methods and parameters in said first partial process and said second partial process, wherein in said database data are available regarding sensor parameters, model spectra, aerosol models, anisotropy types, atmosphere models, together with methods regarding radiative transfer methods, parameterization schemes, assimilation and interpolation methods;
wherein said first partial process is carried out by a first main module (DERA) detecting and identifying dark areas and areas of significant spectral behavior in raw remote sensing data, and said second partial process is carried out by a second main module (CORA) correcting calculations;
said first main module (DERA) consisting of two sub-modules (SSA, GSA), a first one (SSA) of which is responsible for a spectral signature analysis and is used for the identification of reference areas of low reflectance, e.g. of water surfaces and dark forest areas, as well as for the identification of exclusion areas, e.g. clouds and cloud shadows, and a second one (GSA) for a geometric structure analysis, wherein data from different sensors with different geometric and/or spectral resolutions are read as raw data into said first sub-module (SSA), which is aimed at an analysis of the spectral behavior on the pixel level, and processed, and an assignment of the pixels of read-in remotely sensed data set to probably detected remotely sensable objects in reality is made on the basis of generalized model spectra for remotely sensable objects on a pixel level, in such a way that model spectra required for this analysis are stored in the database (DABA), which is accessed during the first partial process, and that in said second sub-module (GSA) of said first main module (DERA) a homogeneity analysis of identified reference areas, an area size analysis, as well as an analysis of the direct and indirect neighborhood is performed in such a way that data areas that were detected by said first sub-module (SSA) are routed to said second sub-module (GSA) and examined for sufficient size; and said second main module (CORA) consisting of two sub-modules (INPRE, DACO), wherein a processing of the required supplementary data is performed in said first sub-module (INPRE) in such a manner that internal supplementary data as they are derived from the raw data and results from said first main module (DERA), are accessed in a first preparation unit (INPRE-INT), and external supplementary data, which are provided via an external interface (EXT-INTF), are processed in a second preparation unit (TNPRE-EXT) in such a manner that standard data assimilation and interpolation methods are used, which are made available in the database (DABA) as methods, and that the actual correction steps are performed in said - 6a -second sub-module (DACO) of said second main module (CORA) with the aid of the supplementary data from said first sub-module (INPRE) in such a way that correction processes are accessed in said method-containing database (DABA).
Pre-tabulated/parameterized radiative transfer calculations make the inventive process fast and, therefore, suitable for operational applications. To attain a good time-space correlation of the atmospheric data with the data in need of correction, these values are estimated, as far as possible, from the data in need of correction. Additional supplementary data that cannot be obtained from the data in need of correction can be acquired externally from operational processing chains via an external interface, and interpolated with suitable methods.
In the numerical process for an automatic atmosphere correction according to the invention, the data from different - 6b -sensors with different geometric and/or spectral resolution may be read in and processed as raw data, e.g, NOAA-AVHRR, ERS-ATSR, (SEA)WIFS, EOS-MODIS, Landsat-TM and Landsat-MSS, IRS-LISS, SPOT-HRV. Essential in the inventive process is the combination of an event-controlled classification and object identification, i.e., a localization and content-based correlation of objects, the actual correction calculation, and the use of current and complete supplementary data regarding the atmospheric condition. Only with this combination can an automatic atmospheric correction take place without interactive intervention or expert knowledge.
The inventive process is, therefore, composed of two partial processes, which are carried out by main modules. The first main. module is used for the detection and identification of dark areas and areas of significant spectral behavior in the remotely sensed data, and the second main module is used for the atmospheric correction of the remotely sensed data.
The two main modules consist of sub-modules. These are supplemented by a database in which basic static and dynamic data, as well as a priori knowledge, e.g., spectral signatures, sensor specifications, statistical properties, correction methods and assimilation methods are stored. This database is accessible by both main modules.
The first main module for its part consists of two sub-modules. The first of these two sub-modules is used for the identification of reference areas of low reflectance, e.g., of water surfaces and dark forest areas, as well as for the identification of exclusion areas, e.g., clouds and cloud shadows. In the process, this sub-module uses the model spectra and sensor-specific information stored in the database. The second one of these two sub-modules is used to perform the homogeneity analysis for identified reference areas (test for representativeness-of the selected areas) and the area size analysis, on one hand, and the analysis of the direct and indirect neighbourhood on the other hand.
The second main module, which is thus used for the atmospheric correction of the remotely sensed data, for its part also consists of two sub-modules. The first one of these two sub-modules is used for processing the required supplementary data. In the process, this sub-module accesses internal supplementary data, as they are derived from the raw data and the results from the first main module, and processes external supplementary data, which are made available via an external interface, e.g., online or via CD-ROM. This may be done using standard data assimilation and interpolation methods, which are made available in the methods database.
_ g _ The second one of these two sub-modules performs the actual correction steps with the aid of the supplementary data-from the first sub-module. In the process, correction methods may be accessed that are stored in the methods database.
BRIEF DESCRIPTION OF THE DRAWINGS
The inventive process for correcting atmospheric influences in multispectral optical remote sensing data will be explained in detail below, based on figure 1 which gives a schematic overview.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS) OF THE
TAT~1G'ATTTIITM
The process shown in the figure for an automatic atmospheric correction according to the invention consists of two different partial processes, which are combined as main modules DERA and CORA. The main module DERA ("Detection of Reference Areas") is used for the detection and identification of dark areas and areas of significant spectral behavior in the remotely sensed data, and the main module CORA
("Correction of Atmosphere") is used for the atmospheric correction of the remotely sensed data. Both main modules DERA and CORA access a database DABA with stored methods and parameters. The process provides the result in the form of a pixel image of the atmosphere-corrected ground reflectance _ g _ with incorporated supplementary atmospheric data,anisotropic reflectance characteristics of different ground type classes, and terrain elevation. Data from different sensors with different geometric and/or spectral resolutions can be read in as raw data and processed.
The main module DERA consists of two sub-modules SSA
("Spectral Signature Analysis") and GSA ("Geometric Structure Analysis"). The spectral signature analysis in the sub-module SSA is aimed at an analysis of the spectral behavior of the multispectral data set on the pixel level. The data from different sensors and/or different geometric resolutions can be read in and processed as raw data. In the process, the pixels of the remotely sensed data set are assigned to the probably detected remotely sensable objects in reality on the basis of generalized model spectra for remotely sensable objects on the pixel level. The model spectra that are required for this analysis are stored in the database DABA, which is used by the algorithm to be executed in the sub-module SSA. Also stored in the database are the sensor specifications that are required~for this analysis, which are made available as a priori knowledge.
On the basis of this information and the spectral behavior of the individual pixels, they are assigned to suitable reference areas, which may serve as dark areas. They furthermore form the basis for the assignment of a ground type class (forest, grassland, bog, savannah, uncultivated field, bushland, water, residential settlement, rock, sand , based on which a suitable function for modeling the anisotropic ground reflectance can be selected in the main module CORA.
The sub-module GSA is used for the homogeneity analysis of identified reference areas, for the area size analysis, as well as for the analysis of the direct and indirect neighbourhood. The data areas detected by the sub-module SSA used for the spectral signature analysis are routed to the sub-module GSA used for the geometric structure analysis, and examined for sufficient size. The geometric structure analysis that is performed in the sub-module GSA, in turn, is divided into a module HMA used for the homogeneity analysis, a module GFA used for the size and shape analysis of identified data areas, and a module NBA used for the neighbourhood analysis.
The module HMA for the homogeneity analysis is used for the identification of contiguous areas, or areas that must be treated separately during the further image processing.
'this is done by examining image areas for their local homogeneity with a filter matrix (m ~ m) and comparing them to threshold values. The measure for the homogeneity, which is then made available as the threshold value for the data set, is derived directly from the data. For this purpose the data set is divided into data sectors and subjected to a statistical analysis with filters of decreasing sizes. The mean value, the standard deviation, as well as the variation coefficient are derived as statistical measures.
In the next module GFA used for the size and shape analysis of the identified data areas, an object identification is assigned to these areas, so that the shape parameters can be directly assigned to the objects. For this purpose, the size of the area and the circumference and compactness of the objects are determined, so that a criterion is derived for the evaluation of the objects as reference areas.
Decision criteria for reference areas are defined as follow-s: reflectance value below a maximum size in the mid-infrared or, alternatively, near-infrared value below a spectral threshold in combination with exceeding a minimum value for a vegetation index, and, additionally, exceeding a minimum area size.
The direct and indirect neighbourhood of the data regions identified as homogenous is subsequently analyzed by the module NBA used for the neighbourhood analysis. The goal of this analysis is to identify so-called mixed pixels and interference pixels, which must be treated separately with respect to their belonging to the adjacent objects to be able to assign to them a suitable ground type class. For this purpose the data are analyzed, with filters of decreasing sizes, with respect to the transition contrasts between the identified data sectors, in order to thus be able to estimate e.g., the influence of clouds or haze and to separate phenomena of different length scales (e. g., small-scale variability of the land use from large-scale change in the atmospheric conditions).
The main module CORA consists of two sub-modules INPRE ("Input Preparation") and DACO ("Data Correction").
The sub-module INPRE is aimed at making the required supplementary data for the sub-module DACO available for each pixel and it is divided into two preparation modules INPRE-INT
and INPRE-EXT. In the process; the module INPRE-INT uses results from the first main module DERA (internal supplementary data); the module INPRE-EXT processes the atmospheric data (external supplementary-data)that are made available via an external interface EXT-IdTF(online or from CD-ROM, e.g., from the German Remote Sensing Data Center, DFD). In the process, the sub-module INPRE determines the following supplementary data for each pixel: turbidity mask, ground mask and exclusion mask (INPRE-INT), trace gas masks and terrain model (INPRE-EXT ) .
With the module INPRE-INT, the aerosol-optical thickness in visible channels above the reference areas determined by the first main module DERA is determined by means of a dark field method and transferred, with suitable interpolation routines, from the database DABA to all pixels and spectral channels (turbidity mask). During the spatial interpolation, the contiguous areas of similar atmospheric conditions that were determined by the first main module DERA
are taken into consideration. The ground type class determined by the first main module DERA is used to select the suitable model function of the anisotropic reflectance properties (ground mask). Cloud and shadow areas determined by the first main module DERA are annotated as "pixels not to be corrected (exclusion mask).
Via the online interface EXT-INTF or via CD-ROM, satellite or airborne data of total ozone_column and water vapor column (e. g., from the German Remote Sensing Data Center, DFD) are loaded by the module INPRE-EXT and converted with suitable assimilation processes (e. g., Harmonic analysis, Kalman filter, Kriging) from the database DABA to the point in time/geographical location of the raw data (trace gas masks).
A suitable section of a digital elevation model is acquired via the same interface and re-projected to the sensor coordinates.
The sub-module DACO performs the pixel-by-pixel correction of all non-excluded pixels (exclusion mask) with the aid of the supplementary data (masks) from the sub-module INPRE. In the process access is made, as desired, to quick correction methods from the database DABA (by means of known radiation transportation programs, e.g., 5S, SOS, MODTRAN
calculated lookup tables, or published parameterization schemes, e.g., SMAC, EXACT).
The sub-module DACO first performs, for each pixel of the raw data, a conversion of the value measured at the top of the atmosphere into a reflectance value on the ground using a module RECD (Reflectance Conversion). The pixel values of the turbidity mask, the ground mask, the trace gas mask and the elevation model are used in the process as supplementary data. The sub-module DACO incorporates the anisotropy of the reflectance from the earth's surface by using a suitable model function for each of the ground types determined from the main module DERA.
For the correction of the incident radiation into the instantaneous field of view of the sensor from adjacent pixels a simple adjacency filter is applied in a module ACO ("Adjacency Correction") using the turbidity mask. This is done in a second step, which, however, is necessary only for high-resolution sensors.
The modules access the methods and parameters database DABA. In this database, data (sensor parameters, model spectra, aerosol models, 'anisotropy types, atmospheric models) are available together with methods ( radiative transfer methods, parameterization schemes, assimilation and interpolation methods).
The raw data must be roughly (approximately ~ 1 pixel or ~ degree) annotated with the geographic position (geographic longitude, geographic latitude) and the observation geometry (observation zenith, observation azimuth) of each individual pixel; only "nearest neighbor" methods should be used as interpolation methods, if need be. The raw data must be multispectral and have at least one visible and one near-infrared channel. The optimum is an additional channel in the medium infrared (more exact dark field method) and a further channel in the visible (more precise spectral interpolation of the aerosol optical thickness.) Together with the observation data the precise observation time must be made available, to be able to calculate the position of the sun.
Alternately, the zenith and azimuth angle of the sun may be provided for each pixel as an additional channel.
Claims (23)
1. A process for correcting atmospheric influences in multispectral optical remote sensing data that are acquired by different satellite or airborne sensors for earth observations with different geometric and/or spectral resolutions, and read in and processed as raw data to generate an image, comprising the steps of:
pre-classification of the raw data for an automatic recognition of pre-defined classes performing in a first partial process; and performing a correction calculation to convert the uncorrected reflectances into corrected reflectances on the ground in a second partial process; and incorporating into the process current and essentially complete supplementary data on the current atmospheric conditions; and accessing a database (DABA) with methods and parameters in said first partial process and said second partial process, wherein in said database data are available regarding sensor parameters, model spectra, aerosol models, anisotropy types, atmosphere models, together with methods regarding radiative transfer methods, parameterization schemes, assimilation and interpolation methods;
wherein said first partial process is carried out by a first main module (DERA) detecting and identifying dark areas and areas of significant spectral behavior in raw remote sensing data, and said second partial process is carried out by a second main module (CORA) correcting calculations;
said first main module (DERA) consisting of two sub-modules (SSA, GSA), a first one (SSA) of which is responsible for a spectral signature analysis and is used for the identification of reference areas of low reflectance, e.g. of water surfaces and dark forest areas, as well as for the identification of exclusion areas, e.g.
clouds and cloud shadows, and a second one (GSA) for a geometric structure analysis, wherein data from different sensors with different geometric and/or spectral resolutions are read as raw data into said first sub-module (SSA), which is aimed at an analysis of the spectral behavior on the pixel level, and processed, and an assignment of the pixels of read-in remotely sensed data set to probably detected remotely sensable objects in reality is made on the basis of generalized model spectra for remotely sensable objects on a pixel level, in such a way that model spectra required for this analysis are stored in the database (DABA), which is accessed during the first partial process, and that in said second sub-module (GSA) of said first main module (DERA) a homogeneity analysis of identified reference areas, an area size analysis, as well as an analysis of the direct and indirect neighborhood is performed in such a way that data areas that were detected by said first sub-module (SSA) are routed to said second sub-module (GSA) and examined for sufficient size; and said second main module (CORA) consisting of two sub-modules (INPRE, DACO), wherein a processing of the required supplementary data is performed in said first sub-module (INPRE) in such a manner that internal supplementary data as they are derived from the raw data and results from said first main module (DERA), are accessed in a first preparation unit (INPRE-INT), and external supplementary data, which are provided via an external interface (EXT-INTF), are processed in a second preparation unit (INPRE-EXT) in such a manner that standard data assimilation and interpolation methods are used, which are made available in the database (DABA) as methods, and that the actual correction steps are performed in said second sub-module (DACO) of said second main module (CORA) with the aid of the supplementary data from said first sub-module (INPRE) in such a way that correction processes are accessed in said method-containing database (DABA).
pre-classification of the raw data for an automatic recognition of pre-defined classes performing in a first partial process; and performing a correction calculation to convert the uncorrected reflectances into corrected reflectances on the ground in a second partial process; and incorporating into the process current and essentially complete supplementary data on the current atmospheric conditions; and accessing a database (DABA) with methods and parameters in said first partial process and said second partial process, wherein in said database data are available regarding sensor parameters, model spectra, aerosol models, anisotropy types, atmosphere models, together with methods regarding radiative transfer methods, parameterization schemes, assimilation and interpolation methods;
wherein said first partial process is carried out by a first main module (DERA) detecting and identifying dark areas and areas of significant spectral behavior in raw remote sensing data, and said second partial process is carried out by a second main module (CORA) correcting calculations;
said first main module (DERA) consisting of two sub-modules (SSA, GSA), a first one (SSA) of which is responsible for a spectral signature analysis and is used for the identification of reference areas of low reflectance, e.g. of water surfaces and dark forest areas, as well as for the identification of exclusion areas, e.g.
clouds and cloud shadows, and a second one (GSA) for a geometric structure analysis, wherein data from different sensors with different geometric and/or spectral resolutions are read as raw data into said first sub-module (SSA), which is aimed at an analysis of the spectral behavior on the pixel level, and processed, and an assignment of the pixels of read-in remotely sensed data set to probably detected remotely sensable objects in reality is made on the basis of generalized model spectra for remotely sensable objects on a pixel level, in such a way that model spectra required for this analysis are stored in the database (DABA), which is accessed during the first partial process, and that in said second sub-module (GSA) of said first main module (DERA) a homogeneity analysis of identified reference areas, an area size analysis, as well as an analysis of the direct and indirect neighborhood is performed in such a way that data areas that were detected by said first sub-module (SSA) are routed to said second sub-module (GSA) and examined for sufficient size; and said second main module (CORA) consisting of two sub-modules (INPRE, DACO), wherein a processing of the required supplementary data is performed in said first sub-module (INPRE) in such a manner that internal supplementary data as they are derived from the raw data and results from said first main module (DERA), are accessed in a first preparation unit (INPRE-INT), and external supplementary data, which are provided via an external interface (EXT-INTF), are processed in a second preparation unit (INPRE-EXT) in such a manner that standard data assimilation and interpolation methods are used, which are made available in the database (DABA) as methods, and that the actual correction steps are performed in said second sub-module (DACO) of said second main module (CORA) with the aid of the supplementary data from said first sub-module (INPRE) in such a way that correction processes are accessed in said method-containing database (DABA).
2. A process according to claim 1, wherein the second sub-module (GSA) of the first main module (DERA) is divided into a homogeneity analysis module (HMA), a size and shaped analysis module (GFA) for identified areas, and an neighbourhood analysis module (NBA), wherein the homogeneity analysis module (HMA) performs an identification of contiguous areas and/or areas that need to be treated separately in the further image processing, that an object identification is assigned to the identified data areas by the size and shape analysis module (GFA) so that the shape parameters can be directly assigned module (NBA) performs an analysis of the direct and indirect neighbourhood of the data regions identified as homogenous, in such a way that the goal of the analysis lies in the identification of so-called mixed and interference pixels that must be treated separately with respect to belonging to the adjacent objects, in order to be able to assign thereto a suitable ground type class.
3. A process according to claim 2, wherein in the homogeneity analysis module (HMA), image areas are examined with a filter matrix (m ~ m) for their local homogeneity and compared to threshold values in such a manner that a measure for the homogeneity, which is then made available as the threshold value for the data set, is derived directly from the data, and that the data set is divided into data sectors for this purpose and statistically analyzed with filters of decreasing size, in such a manner that the mean value, the standard deviation and the variation coefficient are derived as statistical measures.
4. A process according to claim 2, wherein in the size and shape analysis module (GFA), an area size, the circumference and the compactness of the objects are determined, so that a criterion is derived for evaluating the objects as reference areas.
5. A process according to claim 4, wherein the decision criteria for reference areas are defined as the differentiation of a maximum size for the reflectance in the mid-infrared spectrum or, alternately, a near-infrared value below a spectral threshold, in combination with exceeding a minimum value for a vegetation index and, additionally, exceeding a minimum area size.
6. A process according to claim 2, wherein in the adjacency module (NBA), an analysis of the data is performed with filters of decreasing size with respect to the transition contrasts between identified data sectors, in order to thus be able to estimate the influence of clouds or haze, and to be able to separate phenomena of different length scales, e.g., small-scale variability of the land use from large-scale changes in the atmospheric conditions.
7. A process according to claim 1, wherein the first sub-module (INPRE) of the second main module (CORA), a turbidity mask, a ground mask and an exclusion mask are determined for each pixel in the first preparation unit (INPRE-INT), and trace gas masks and a terrain model are determined in the second preparation unit (INPRE-EXT).
8. A process according to claim 7, wherein with the first preparation unit (INPRE-INT), an aerosol-optical thickness in visible channels above the reference areas determined by the first main module (DERA) is determined by means of a dark field method and transferred to all pixels and spectral channels with suitable interpolation routines from the database (DABA) to form the turbidity mask, and that the contiguous areas of similar atmospheric conditions determined by the first main module (DERA) are incorporated into the spatial interpolation.
9. A process according to claim 7, wherein a ground type class determined by the first main module (DERA) is used in the first preparation unit (INPRE-INT) for the selection of a suitable model function of the anisotropic reflectance characteristics in the generation of a ground mask.
10. A process according to claim 7, wherein cloud and shadow areas determined by the first main module (DERA) are annotated in the first preparation unit (INPRE-INT) during the generation of the exclusion mask as pixels that are not to be corrected.
11. A process according to claim 7, wherein for generation of the trace gas masks, satellite or airborne data of total ozone column and water vapor column are loaded by the second preparation unit (INPRE-EXT) via an interface (EXT-INTF) e.g., an online interface or via CD-ROM, and converted to the time/geographic location of the raw data by means of suitable assimilation methods, such as Harmonic analysis, Kalman filter or Kriging from the database (DABA).
12. A process according to claim 7, wherein via an interface that is or may be identical to the interface of claim 16, a suitable section of a digital elevation model is acquired by the second preparation unit (INPRE-EXT) and re-projected to the sensor coordinates.
13. A process according to claim 1, wherein in the second sub-module (DACO) of the second main module (CORA), the correction of all pixels not excluded by the exclusion mask is performed pixel by pixel with the aid of the supplementary data from the first sub-module (INPRE) of the second main module (CORA).
14. A process according to claim 13, wherein access is made, as desired, during correction to fast correction methods from the database (DABA), particularly to know radiation transportation programs, such as 6S, SOS, MODTRAN- calculated lookup table or published parameterization schemes, e.g., SMAC, EXACT.
15. A process according to claim 14, wherein in the second sub-module (DACO) of the second main module (CORA), a conversion (RECO) of the value measured at the top of the atmosphere is performed for each pixel of the raw data into a reflectance value on the ground in such a manner that the pixel values of the turbidity mask, the ground mask, the trace gas mask and the elevation model are used as supplementary data, and that the anisotropy of the reflectance from the ground is incorporated by the second sub-module (DACO) of the second main module (CORA) by using a suitable model function for each ground type determined from the first main module (DERA).
16. A process according to claim 13, wherein when high-resolution sensors are used, a simple filter using the turbidity mask is additionally used, in a second step, in the second sub-module (DACO) of the second main module (CORA) for correction of incident radiation into the instantaneous field of view of a sensor from adjacent pixels.
17. A process according to claim 7, wherein a pixel image of the-atmosphere-corrected ground reflectance is obtained with incorporated supplementary atmospheric data, anisotropic reflectance characteristics of various ground type classes, and terrain elevation.
18. A process according to claim 7, wherein the raw data are roughly (approximately ~1 pixel or ~1 degree) annotated with the geographic position (geographic longitude, geographic latitude) and the observation geometry (observation zenith, observation azimuth) of each individual pixel.
19. A process according to claim 7, wherein the multispectral raw data have at least one visible and one nearinfrared spectrum.
20. A process according to claim 19, wherein one additional channel in the middle infrared spectrum is provided for a more precise dark field method, and a further channel in the visible light spectrum is provided for a more precise spectral interpolation of the aerosol optical thickness.
21. A process according to claim 1, wherein an exact time an observation is made available with the observation data for the calculation of the given position of the sun.
22. A process according to claim 7, wherein a current zenith and azimuth angle of the sun is provided for each pixel as an additional channel.
23. A process according to claim 1, wherein atmospheric influences in multispectral thermal remote sensing data are corrected.
List of Reference Numerals ACO Adjacency filter CORA Main module for atmospheric correction DABA Database DACO Sub-module for data correction DERA Main module for detection of reference areas EXT-INF External interface GFA Module for size and shape analysis GSA Sub-module for geometric structure analysis HMA Module for homogeneity analysis INPRE Sub-module for data preparation INPRE-INT Preparation module for internal data INPRE-EXT Preparation module for external data NBA Module for adjacency analysis RECO Model for reflectance coefficient conversion SSA Sub-module for spectral signature analysis
List of Reference Numerals ACO Adjacency filter CORA Main module for atmospheric correction DABA Database DACO Sub-module for data correction DERA Main module for detection of reference areas EXT-INF External interface GFA Module for size and shape analysis GSA Sub-module for geometric structure analysis HMA Module for homogeneity analysis INPRE Sub-module for data preparation INPRE-INT Preparation module for internal data INPRE-EXT Preparation module for external data NBA Module for adjacency analysis RECO Model for reflectance coefficient conversion SSA Sub-module for spectral signature analysis
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP99113927.0 | 1999-07-16 | ||
| EP99113927A EP1091188B1 (en) | 1999-07-16 | 1999-07-16 | Method for correcting atmospheric influences in multispectral optical teledetection data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CA2313812A1 CA2313812A1 (en) | 2001-01-16 |
| CA2313812C true CA2313812C (en) | 2007-01-16 |
Family
ID=8238609
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA002313812A Expired - Fee Related CA2313812C (en) | 1999-07-16 | 2000-07-12 | Process for correcting atmospheric influences in multispectral optical remote sensing data |
Country Status (4)
| Country | Link |
|---|---|
| US (1) | US6484099B1 (en) |
| EP (1) | EP1091188B1 (en) |
| CA (1) | CA2313812C (en) |
| DE (1) | DE59910670D1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102507474A (en) * | 2011-10-28 | 2012-06-20 | 大连海事大学 | Method and system for identifying oil spilling target of ship |
Families Citing this family (39)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6683970B1 (en) * | 1999-08-10 | 2004-01-27 | Satake Corporation | Method of diagnosing nutritious condition of crop in plant field |
| US7337065B2 (en) * | 2001-01-23 | 2008-02-26 | Spectral Sciences, Inc. | Methods for atmospheric correction of solar-wavelength hyperspectral imagery over land |
| US6714868B2 (en) * | 2001-06-28 | 2004-03-30 | The United States Of America As Represented By The Secretary Of The Navy | Similarity transformation method for data processing and visualization |
| US7043369B2 (en) * | 2003-04-15 | 2006-05-09 | The Johns Hopkins University | Radiance library forecasting for time-critical hyperspectral target detection systems |
| DE10358938B4 (en) * | 2003-12-15 | 2006-06-29 | Deutsches Zentrum für Luft- und Raumfahrt e.V. | Method and device for the at least partially automated evaluation of remote sensing data |
| DE102004024595B3 (en) * | 2004-04-30 | 2005-08-18 | Deutsches Zentrum für Luft- und Raumfahrt e.V. | Determining usability of remote sensing data involves determining distances between first and second image matrix elements and automatically assessing usability of image matrix using the distances |
| ATE443846T1 (en) | 2004-09-15 | 2009-10-15 | Deutsch Zentr Luft & Raumfahrt | PROCESSING OF REMOTE SENSING DATA |
| DE102004052603B3 (en) | 2004-10-29 | 2006-05-24 | Deutsches Zentrum für Luft- und Raumfahrt e.V. | Method for eliminating shadow effects in remote sensing data over land |
| DE102004057855A1 (en) * | 2004-11-30 | 2006-06-08 | Deutsches Zentrum für Luft- und Raumfahrt e.V. | Method of detecting hot targets on the ground |
| US7680337B2 (en) * | 2005-02-22 | 2010-03-16 | Spectral Sciences, Inc. | Process for finding endmembers in a data set |
| US8189877B2 (en) | 2005-10-21 | 2012-05-29 | Carnegie Institution Of Washington | Remote sensing analysis of forest disturbances |
| GB2435523B (en) | 2006-01-25 | 2010-06-23 | Arkex Ltd | Terrain correction systems |
| US7558673B1 (en) * | 2006-03-03 | 2009-07-07 | Itt Manufacturing Enterprises, Inc. | Method and system for determining atmospheric profiles using a physical retrieval algorithm |
| US8478067B2 (en) * | 2009-01-27 | 2013-07-02 | Harris Corporation | Processing of remotely acquired imaging data including moving objects |
| US8260086B2 (en) * | 2009-03-06 | 2012-09-04 | Harris Corporation | System and method for fusion of image pairs utilizing atmospheric and solar illumination modeling |
| US9097792B2 (en) | 2009-08-12 | 2015-08-04 | The Johns Hopkins University | System and method for atmospheric correction of information |
| CN102033898B (en) * | 2010-09-27 | 2012-05-23 | 华东师范大学 | Extraction method for local cloud cover information metadata of moderate resolution imaging spectral image |
| US9576349B2 (en) | 2010-12-20 | 2017-02-21 | Microsoft Technology Licensing, Llc | Techniques for atmospheric and solar correction of aerial images |
| US8666190B1 (en) | 2011-02-01 | 2014-03-04 | Google Inc. | Local black points in aerial imagery |
| CN102288956B (en) * | 2011-05-10 | 2013-04-03 | 中国资源卫星应用中心 | Atmospheric correction method for multispectral data of remote sensing satellite |
| CN102279393B (en) * | 2011-07-15 | 2013-05-22 | 北京航空航天大学 | Cross radiometric calibration method of hyper-spectral sensor based on multi-spectral sensor |
| CN102495405B (en) * | 2011-11-30 | 2013-06-19 | 武汉大学 | Evaluation method of TM/ETM (thematic mapper/enhanced thematic mapper) and image-based atmospheric correction product quality |
| US8693938B2 (en) * | 2012-03-02 | 2014-04-08 | Xerox Corporation | Apparatus and systems for high pressure fusing electrostatic offset mitigation |
| CN102955878B (en) * | 2012-09-05 | 2015-07-29 | 环境保护部卫星环境应用中心 | Based on the Inland Water optics sorting technique of MERIS full resolution image data |
| CN103018403B (en) * | 2012-11-23 | 2015-03-04 | 河南科技大学 | Method for improving dynamic measurement for vehicle-mounted system based on composite dimension |
| CN103968808B (en) * | 2013-01-24 | 2016-12-28 | 国家基础地理信息中心 | The strict geometric correction method of wide visual field Satellite CCD image |
| US20140270502A1 (en) * | 2013-03-15 | 2014-09-18 | Digitalglobe, Inc. | Modeled atmospheric correction objects |
| US9396528B2 (en) * | 2013-03-15 | 2016-07-19 | Digitalglobe, Inc. | Atmospheric compensation in satellite imagery |
| US9449244B2 (en) * | 2013-12-11 | 2016-09-20 | Her Majesty The Queen In Right Of Canada, As Represented By The Minister Of National Defense | Methods for in-scene atmospheric compensation by endmember matching |
| CN104156567B (en) * | 2014-07-23 | 2017-05-03 | 中国科学院遥感与数字地球研究所 | Technique for acquiring surface reflectance by coupling satellite remote-sensing image atmospheric correction and topographical correction processes |
| CN106875365B (en) * | 2017-03-01 | 2019-06-11 | 淮安信息职业技术学院 | Spill image partition method based on GSA |
| US10820472B2 (en) | 2018-09-18 | 2020-11-03 | Cnh Industrial America Llc | System and method for determining soil parameters of a field at a selected planting depth during agricultural operations |
| CN109556715A (en) * | 2018-11-19 | 2019-04-02 | 中国国土资源航空物探遥感中心 | A kind of more air strips image Radiometric Correction Methods of Airborne Hyperspectral |
| CN111325184B (en) * | 2020-03-20 | 2023-04-18 | 宁夏回族自治区自然资源勘测调查院 | Intelligent interpretation and change information detection method for remote sensing image |
| CN113466917B (en) * | 2020-03-31 | 2023-12-15 | 中国矿业大学(北京) | Forced calculation method for soot type aerosol radiation |
| CN111398180B (en) * | 2020-04-28 | 2023-04-18 | 中国交通通信信息中心 | Atmospheric compensation method suitable for thermal infrared remote sensing image |
| KR102500371B1 (en) * | 2022-11-03 | 2023-02-16 | 대한민국(관리부서: 환경부 국립환경과학원장) | Aerosol Mass Flow Rate Expression System Through Environmental Satellite And Its Method |
| CN116879237B (en) * | 2023-09-04 | 2023-12-12 | 自然资源部第二海洋研究所 | Atmospheric correction method for offshore turbid water body |
| CN117436286B (en) * | 2023-12-20 | 2024-03-26 | 中国科学院合肥物质科学研究院 | Aerosol and water vapor inversion and aerial remote sensing image correction method, device and equipment |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE4102579A1 (en) * | 1990-01-30 | 1991-08-01 | Deutsche Forsch Luft Raumfahrt | Compensating atmospheric effects in solar and thermal spectral range - filtering image and compensation functions recalled from memory for images obtained by opto-electronic satellite sensors |
| US5324113A (en) * | 1992-12-10 | 1994-06-28 | E-Systems, Inc. | Process for multispectral/multilook atmospheric estimation |
| US6356646B1 (en) * | 1999-02-19 | 2002-03-12 | Clyde H. Spencer | Method for creating thematic maps using segmentation of ternary diagrams |
-
1999
- 1999-07-16 EP EP99113927A patent/EP1091188B1/en not_active Expired - Lifetime
- 1999-07-16 DE DE59910670T patent/DE59910670D1/en not_active Expired - Lifetime
-
2000
- 2000-07-12 CA CA002313812A patent/CA2313812C/en not_active Expired - Fee Related
- 2000-07-14 US US09/617,542 patent/US6484099B1/en not_active Expired - Fee Related
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102507474A (en) * | 2011-10-28 | 2012-06-20 | 大连海事大学 | Method and system for identifying oil spilling target of ship |
Also Published As
| Publication number | Publication date |
|---|---|
| DE59910670D1 (en) | 2004-11-04 |
| EP1091188B1 (en) | 2004-09-29 |
| US6484099B1 (en) | 2002-11-19 |
| CA2313812A1 (en) | 2001-01-16 |
| EP1091188A1 (en) | 2001-04-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2313812C (en) | Process for correcting atmospheric influences in multispectral optical remote sensing data | |
| Du et al. | Retrieval of global terrestrial solar-induced chlorophyll fluorescence from TanSat satellite | |
| Du et al. | Radiometric normalization, compositing, and quality control for satellite high resolution image mosaics over large areas | |
| Huang et al. | Development of time series stacks of Landsat images for reconstructing forest disturbance history | |
| Hall et al. | Radiometric rectification: toward a common radiometric response among multidate, multisensor images | |
| Schmidt et al. | Spectral discrimination of vegetation types in a coastal wetland | |
| Levy et al. | Algorithm for remote sensing of tropospheric aerosol over dark targets from MODIS: Collections 005 and 051: Revision 2; Feb 2009 | |
| Remer et al. | Algorithm for remote sensing of tropospheric aerosol from MODIS: Collection 005 | |
| Gong et al. | Estimation of forest leaf area index using vegetation indices derived from Hyperion hyperspectral data | |
| Teillet et al. | Effects of spectral, spatial, and radiometric characteristics on remote sensing vegetation indices of forested regions | |
| Theseira et al. | An evaluation of spectral mixture modelling applied to a semi-arid environment | |
| Coops et al. | Chlorophyll content in eucalypt vegetation at the leaf and canopy scales as derived from high resolution spectral data | |
| CN108932521B (en) | Deep learning-based crop classification method and system | |
| Rautiainen et al. | Coupling forest canopy and understory reflectance in the Arctic latitudes of Finland | |
| Manakos et al. | Comparison between atmospheric correction modules on the basis of worldview-2 imagery and in situ spectroradiometric measurements | |
| Panigada et al. | Chlorophyll concentration mapping with MIVIS data to assess crown discoloration in the Ticino Park oak forest | |
| Seixas et al. | Comparative analysis of MODIS-FAPAR and MERIS–MGVI datasets: Potential impacts on ecosystem modeling | |
| Heiden et al. | Ecological evaluation of urban biotope types using airborne hyperspectral hymap data | |
| Darvishsefat et al. | Application of hyperspectral data for forest stand mapping | |
| Kawishwar | Atmospheric correction models for retrievals of calibrated spectral profiles from hyperion EO-1 data | |
| GASTELLU-ETCHEGORRY et al. | Computer-assisted land cover mapping with SPOT in Indonesia | |
| Ibrahim et al. | Assessing the discrepancy in open-source atmospheric correction of Sentinel-2 acquisitions for a tropical mining area in New Caledonia | |
| RU2018156C1 (en) | Method for detecting polluted areas in atmosphere | |
| Wilson et al. | Automated selection of suitable atmospheric calibration sites for satellite imagery | |
| Nadzirah et al. | Application of Sentinel-2A Images for Land Cover Classification Using NDVI in Jember Regency |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request | ||
| MKLA | Lapsed |