|Publication number||USRE41364 E1|
|Application number||US 11/433,909|
|Publication date||Jun 1, 2010|
|Filing date||May 11, 2006|
|Priority date||Feb 18, 2000|
|Also published as||US6734998, US20010019431|
|Publication number||11433909, 433909, US RE41364 E1, US RE41364E1, US-E1-RE41364, USRE41364 E1, USRE41364E1|
|Inventors||Yu-Fen Tsai, Te-Chih Chang|
|Original Assignee||Yu-Fen Tsai, Te-Chih Chang|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (10), Classifications (31), Legal Events (3)|
|External Links: USPTO, USPTO Assignment, Espacenet|
1. Field of the Invention
The present invention provides a method for determining if an image from a scanner has occurrences of scan line misalignments. More particularly, a software method enabling a program to determine if an image from a scanner has occurrences of scan line misalignments is disclosed.
2. Description of the Prior Art
Scanners are popular computer peripherals that are used to digitize documents or pictures so that they may be stored on a computer. To ensure a high quality of these scanned images, manufacturers strive to increase the resolution of the images, and to make their colors more brilliant. But a key factor affecting the quality of scanned images is the stability of the scanning module. If the stability of the scanning module is insufficient, the images from a scanner may have misalignments or entire deletions of scan lines in the image.
Please refer to FIG. 1.
Please refer to FIG. 2.
In this prior art, a search is performed within the scanned test image 20 to find the positions of the boundary points of the test image 20, and then pixel values within the boundary points are tested against diagonally adjacent pixel values. For example, I(X, Y) represents the pixel value of the test image 20 at the Xth column and the Yth line. A simple program is used to compare the pixel value I(i) of a point (X, Y) and the pixel value I(i+1) of another point (X−1, Y+1). If the difference between I(i) and I(i+1) is too large, then it is assumed that a scan line 22 is missing between the lines (Y) and (Y+1).
Hence, the prior art compares two adjacent lines and determines if the scanned test image 20 conforms to the expected 45 degree symmetry of the test picture. The minimum unit required to determine if a scan line has been skipped is one pixel. This is not accurate enough to satisfy the requirements of a high-end scanner.
It is therefore an objective of the present invention to provide a method for determining scan line misalignments of a scanner.
Briefly, the present invention scans a test image that has a black bias on a white background. The black bias is a line set at about 45 degrees to the scan lines of the scanner. The method involves finding boundary points of the scanned bias, calculating a regression line from the positions of the boundary points, using differences in the position of adjacent boundary points together with the slope reciprocal of the regression line to determine error values, and comparing the error values with a gate value to determine if there are any occurrences of scan line misalignment.
It is an advantage that the present invention can detect scan line misalignments with sub-pixel accuracy, thus fulfilling the more rigid requirements for high-level scanners.
These and other objectives of the present invention will no doubt become obvious to these of ordinary skill in the art after reading the following detailed description of the preferred embodiment, which is illustrated in the various figures and drawings.
Please refer to FIG. 3 and FIG. 4.
To determine if there are any occurrences of scan line misalignment, the scanning module 38 is used to scan the document 36. Scan line image information for a plurality of scan lines is collected, each scan line containing a portion of scanned image of the document 36, the scan line information being collected in order. The image information in each scan line includes a plurality of gray-scale pixels, a portion of which correspond to the bias 37.
A searching method is used to find a boundary point of the bias 37 from the gray-scale image information in each scan line. Because the bias 37 on the document 36 includes two boundary lines, the image information in each scan line will have two boundary points. For convenience, the positions of the boundary points of the left boundary line will be described. The positions of the boundary points on the right side of the bias 37 are found in the same manner. This method is actually quite well known in the field of image processing.
Please refer to FIG. 5.
Moving forward from the first boundary reference point P1 by a first predetermined number (say, 15) of pixels, a second predetermined number of pixels (again, 15) are selected. Average of the gray-scale values of these chosen pixels are used to define a white reference level (VW). Similarly, by moving backwards from the first boundary reference point P1 by the first predetermined number of pixels (15), another group of the second predetermined number of pixels (15) are selected, and their average values are used as a black reference level (VB).
Because the first boundary reference point P1 is located in an interim region 46 between the white region 42 and the black region 44, moving forward or backward by a predetermined number of pixels from the first boundary point P1 is used to ensure that the chosen pixels are located in the interim region 46. This makes the calculation of the white and the black reference levels more accurate.
The average of the white reference level (VW) and the black reference level (VB) is used to define a boundary level (V0). Two adjacent pixels P2 and P3 are then chosen between where the boundary level V0 falls. That is, the gray-scale image value of the boundary level V0 lies between the gray-scale image values of the pixels P2 and P3, satisfying the inequality VP3≦V0≦VP2. These two pixels are used as a second and a third boundary reference points P2 and P3.
Please refer to FIG. 6.
Because P2 and P3 are adjacent pixels, P3 is equal to P2+1. From the equation
the position of the boundary point X is equal to
In the same manner, a series of different positions of boundary points Xi can be calculated, where i is an integer ranging from 1 to n, n being the number of scan lines.
After determining the boundary position points of the left boundary line for every scan line, a regression line and its slope can be calculated. The calculation of the regression line can be done using all, or only some, of the previously found boundary points. Because the calculation and mathematical significance of the regression line is a well known prior art mathematical tool, the equations are only noted here, without undue explanation. With n parts of numbers (xi, yi), i from 1 to n and being a positive integer, chosen to calculate the regression line y=mx+b, the values of m and b can be determined from the following equations:
After calculating the regression line, the difference in the position of every boundary point is calculated (Δxi=xi−xi+1). This, with the reciprocal of the slope of the regression line (1/m), is used to calculate corresponding error values as (ERRi=|Δxi−1/m|). And the error values can also be interpreted with the use of error ratios, which are equal to ERRi/(1/m).
Finally, each error value ERRi is compared against a predetermined gate value (TD) to determine if the image from the scanner has any occurrences of scan line misalignment. If a specific error value is larger than the gate value, then there must be a scan line misalignment at the corresponding scan line. If the specific error value is less than or equal to the gate value, then there is no occurrence of scan line misalignment at that scan line. Occurrences of scan line misalignment are therefore determined by the choice of the gate value, which must be set by experienced personnel. In the preferred embodiment of the present invention, the gate value is about 0.3.
Please refer to FIG. 7.
Please refer to FIG. 8.
In the first prior art method, a manual, visual inspection of a scanned test image is performed, which is both judgmental (as it depends on the individual experience of the testing staff) and time-consuming. In the second prior art, the comparison of gray-scale image values of two adjacent scan lines to determine if the slope of the scanned test image is equal to its original value lowers the factors of personal judgement, but the results are too rough to fulfill the requirements of high-level scanners. The present invention method, however, searches for the boundary points to determine a regression line, and can calculate the positions of the boundary points accurately within one pixel unit and the error values of all boundary points to the regression line. If the error value is larger than a predetermined gate value, a scan line misalignment is determined to have occurred. In light of the discussion above, this method fulfills the requirements of high-level scanners, and the gate value can be chosen by experienced personnel to account for different requirements.
Those skilled in the art will readily observe that numerous modifications and alternations of the device may be made while retaining the teachings of the invention. Accordingly, the above disclosure should be construed as limited only by the metes and bounds of the appended claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5563403 *||Dec 14, 1994||Oct 8, 1996||Ricoh Co., Ltd.||Method and apparatus for detection of a skew angle of a document image using a regression coefficient|
|US5892854 *||Jan 21, 1997||Apr 6, 1999||Xerox Corporation||Automatic image registration using binary moments|
|US5978102 *||Nov 22, 1996||Nov 2, 1999||Minolta Co., Ltd.||Image reading apparatus|
|US6134028 *||Jul 6, 1998||Oct 17, 2000||Samsung Electronics Co., Ltd.||Method for scanning documents|
|US6178015 *||Jun 5, 1998||Jan 23, 2001||Mustek Systems, Inc.||Apparatus and method for increasing the scan accuracy and quality of the flatbed scanner by using close loop control|
|US6303921 *||Nov 23, 1999||Oct 16, 2001||Hewlett-Packard Company||Method and system for capturing large format documents using a portable hand-held scanner|
|US6512585 *||Jan 19, 2000||Jan 28, 2003||Murata Manufacturing Co. Ltd||Laser scanning position detecting device|
|US6534757 *||Jan 25, 1999||Mar 18, 2003||Canon Kabushiki Kaisha||Image sensor noise reduction|
|US20020070331 *||Jan 25, 1999||Jun 13, 2002||Fumihiro Inui||Image sensor|
|US20030020821 *||Jul 24, 2002||Jan 30, 2003||Tohru Watanabe||Imaging apparatus|
|U.S. Classification||358/488, 358/494, 358/1.9, 382/291, 358/486, 358/406, 250/208.1, 358/474, 348/372, 358/409, 382/287, 348/241|
|International Classification||H04N1/00, H04N1/10, H04N1/193, H04N1/04|
|Cooperative Classification||H04N1/1017, H04N1/00053, H04N1/00045, H04N1/193, H04N1/00063, H04N1/00002, H04N1/00018, H04N1/00031, H04N2201/04703|
|European Classification||H04N1/00A3C, H04N1/00A2E, H04N1/00A3J, H04N1/00A3M, H04N1/00A3T, H04N1/00A|
|Jan 11, 2007||AS||Assignment|
Owner name: MUSTEK SYSTEMS, INC.,TAIWAN
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:TSAI, YU-FEN;CHANG, TE-CHIH;REEL/FRAME:018749/0826
Effective date: 20001218
Owner name: TRANSPACIFIC OPTICS LLC,DELAWARE
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:MUSTEK SYSTEMS, INC.;REEL/FRAME:018749/0835
Effective date: 20051202
|Sep 23, 2011||FPAY||Fee payment|
Year of fee payment: 8
|Oct 27, 2015||FPAY||Fee payment|
Year of fee payment: 12