US 20050200714 A1
A digital video system including a computer connected via a network to a number of video servers and cameras. The computer includes a program that provides a grid of display windows, each of which displays an image received from the camera associated with that window. The program sequentially polls each camera, accessing and displaying an image from the camera in its associated window. The program can access the cameras at different frame rates. The program stores image streams in a single file, concatenating each successive image onto the end of the file. The file is then indexed using SOI and EOI markers to permit fast access to individual images within the file. The program can monitor received video and automatically start recording upon detecting motion within the video stream. Motion detection is implemented by comparing color component values for pixels from different images.
1. A method of controlling access by a computer program to a video server, comprising the steps of:
accessing a video server over a network from a computer using a computer program running on the computer;
obtaining from the video server a unique number stored in the video server;
retrieving at least one authorized unique number;
determining if the computer program is authorized to access the video server by comparing the unique number from the video server to the one or more authorized unique numbers and, if access is authorized, then initiating a video stream from the video server to the computer over the network.
2. A method of controlling access by a computer program to a hardware device, comprising the steps of:
providing a computer program stored in computer readable form at a computer, the computer being connected to a hardware device via a network;
accessing the hardware device from the computer over the network using an IP address associated with the hardware device;
obtaining a MAC address from the hardware device over the network;
determining if the program is authorized to access the hardware device by validating the MAC address using data accessible to the client computer; and
if program is authorized to access the hardware device, then the method further comprises the step of permitting the computer program to retrieve data from the hardware device over the network using the IP address.
3. A method of licensing a computer program on a per-device basis to prevent use of the computer program with unauthorized hardware devices, comprising the steps of:
identifying one or more authorized hardware devices, each having a MAC address;
providing a computer program in computer readable form, wherein the computer program is operable upon execution by a computer to permit a user of the computer to input network address information concerning a specific hardware device accessible to the computer via a network;
embedding the MAC address(es) of the authorized hardware device(s) into the computer program; and
executing the computer program on the computer by a user and providing access by the user to the specific hardware device if the specific hardware device has a MAC address that matches one of the MAC addresses embedded into the computer program.
This application is a continuation of U.S. application Ser. No. 09/808,543, filed Mar. 14, 2001, now U.S. Pat. No. ______, which claims the priority of U.S. Provisional Application Ser. No. 60/189,162, filed Mar. 14, 2000. The entire disclosure of U.S. application Ser. No. 09/808,543 is hereby incorporated by reference.
This invention relates to systems for accessing, recording, and displaying camera images from any of a number of remotely located cameras and, more particularly, to such systems that provide access to images from one or more remote cameras over a public or private computer network.
With the continuing expansion and availability of public and private computer networks it is becoming increasingly common to use these networks for remote video and image surveillance. Historically, analog systems have been used for CCTV systems for purposes such as surveillance. They include an analog video camera, a video cable, and a monitor/TV and/or a VCR recording device. Multiple cameras can be hooked to multiple recording VCRs for complete coverage of one or more target areas. Specialized equipment known as a multiplexer or ‘MUX’ can be used to allow multiple cameras to be hooked to a single recording/view device. The multiplexer takes all video feeds in a sequential fashion, recording from one camera at a time. This allows the quantity of cameras to share a single recording tape. Besides the limitations of a serial tape system, additional limitations are introduced when the sequencing rate of the multiplexer is too slow to allow sufficient video record/view speed of a given camera view. Multiplexers are typically expensive and have limited expandability without purchasing additional complete systems. Also, the configuration of these systems usually requires a skilled technician to be available at the systems site which increases the total cost of implementing such systems.
Digital systems have become more prevalent with the advent of standardized digital componentry. These systems utilize the same analog cameras and cabling, but introduce a capture card based collector. This collector can be a proprietary digital device or a PC based computer, either of which has analog video inputs directly connected to it. These inputs convert the video to digital for viewing and recording and may even retransmit the signal to analog tape for recording purposes. A factor limiting these digital systems is that an autonomous computer is required relatively near the video sources, sometimes with user intervention required at regular intervals. These machines are also hardware bound. The number of capture cards is limited to the specific design of the collecting equipment. These systems are usually proprietary to a particular manufacturer and can be very expensive. Remote viewing is usually not available. Failover or redundancy function is also limited and expensive, due to the fact that the system is usually duplicated for redundancy. As with the analog systems discussed above, these systems also have configuration requirements that typically require a skilled technician.
There now exists commercially available networkable cameras that can be accessed over networks running TCP/IP, including both LANs and global networks such as the Internet. Ethernet-based digital video servers are now common that are small, autonomous, and usually contain a web-based configuration utility, as well as administration software. These cameras can be accessed and, in the case of pan/tilt/zoom (PTZ) cameras, controlled over the network using an assigned IP address and standard CGI-based URL syntax or other manufacturer-specified addressing protocols. This allows an authorized user to control the product from anywhere via the Internet or a dialup connection, and allows live images and image streams (video) to be accessed remotely using standard web browsers.
The video servers exist in two forms. One is a camera server that is a complete product containing both a camera and a web server with an Ethernet port. The other is a component based video server with inputs for one or more analog video feeds, which the user can connect to conventional camera PAL or NTSC video feeds. The inputted analog video feeds are converted to digital signals and sent from the video servers' Ethernet port. Thus, the video servers (whether integrated in as part of a camera server or as a standalone unit) can be connected to the Ethernet-based networks commonly used in businesses and other computer enabled sites. These video servers can be connected to these network segments and are fully compatible with existing data on these networks. The video data can be received by standard PC computers which require no special hardware other than an Ethernet connection. The cameras can be easily configured by a novice user who has very basic experience with the Internet.
Ethernet video servers connect to an Ethernet connection and deliver digital video based on user requests or internal scripting agents. A user requests video images via standard CGI enhanced URL syntaxes. These syntaxes control the image metrics and other features of the requested video stream. The images are sent to the user as either static JPG snapshots, or as continuous JPG streams. Rates to 30 FPS are easily attainable. Since these images are delivered by Ethernet, the camera servers are very robust. Although a requested image may not be received completely in an expected time frame, the video server will wait for the user to complete its requests and processing. This virtually guarantees delivery of video, except where a connection to the video server is terminated.
When the user requests a video image or stream, the user is actually requesting a static image that appears to exist as a file in a directory structure on the video server. When the user requests a copy of this image, the video server actually updates it with a new image from the camera source, and the user receives a picture that is up to date. Subsequent requests are to the identical file name, and the server does the updating of its content.
Although IP-based network cameras and camera servers have now evolved to a relatively advanced state, the use of a browser-based interface to this hardware has seemingly impeded development of user interfaces that provide simplified, automated control over the acquisition of snapshot and streaming images over the network. Access to the camera images typically requires knowledge of the manufacturer's CGI-based syntax to access snapshot or streaming images. For example, to access a particular camera, the user may have to specify to the browser an address in the form of http//Uid:PW@18.104.22.168/cgi-bin/fullsize.jpg?camera=1&compression=1. While this may be handled easily enough for a single camera by bookmarking or pulling the URL out of the browser's history buffer, the task becomes more difficult when the user desires to change the access parameters or where different cameras need to be accessed. Where the identified camera cannot be accessed, such as for example due to an improper address being specified, the user may simply receive a standard “404 not found” error message that is not helpful in diagnosing either the error or actual reason why access was not available.
Moreover, browser-based access is typically limited to either a snapshot mode or streaming images. In the snapshot mode, a single image is returned when the appropriate URL is entered into the browser. Subsequent images from the camera are then accessed using the browser's “reload” or “refresh” button. In the streaming mode, once the appropriate URL is specified, the remote server or camera simply begins streaming image files back to the browser. This results in relatively high network utilization that may be undesirable in a shared network environment.
It is therefore a general object of this invention to provide an improved user interface and approach to the network transmission of images from commercially available network cameras.
In accordance with the invention, there is provided a digital video system which includes a client computer, one or more video servers accessible by the client computer over a network, and a plurality of cameras connected to the video servers. The one or more video servers provide an interface between the cameras and client computer, with images from the cameras being accessed by the video servers and then sent to the client computer upon request by the client computer. The client computer provides a user interface display on the computer that includes a display window for each of the cameras accessed over the network and that displays in each of the display windows an image received from the camera associated with that display window.
In accordance with another aspect of the invention, there is provided a computer readable medium for use by the client computer in providing the user interface display. The computer readable medium includes a digital storage device and a user interface program stored on the digital storage device in computer readable form. The program is operable upon execution by the computer to access camera data uniquely identifying the cameras and to attempt access to the cameras over the network using the camera data. The program is further operable to verify access to those ones of the cameras that are accessible over the network and to generate the user interface display and to display in each of the display windows an image received from the camera associated with that display window.
In accordance with another aspect of the invention, there is provided a computer readable medium and method for use by a computer in detecting motion in a sequence of color video images received by the computer. The computer readable medium includes a digital storage device containing a program that is operable to carry out the motion detection. The program selects one of the video images as a reference image and another of the video images as a selected image to be compared to the reference image. The program then accesses a color component value for each of a number of pixels from the reference image and for each of a number of corresponding pixels from the selected image, with the program being operable to perform pixel comparisons using the color component values and to generate a motion detect signal when the difference in color component values for corresponding pixels from the reference and selected images differ by more than a preselected amount.
In accordance with yet another aspect of the invention, there is provided a computer readable medium and method for use by a computer in recording a sequence of images, each of which comprises an individual image file in which the image is represented in a graphics file format that includes a first marker identifying the start of the image and a second marker identifying the end of the image. The computer readable medium includes a digital storage device containing a program that is operable to carry out the recording. The program stores the images together as a single file that comprises the group of images concatenated together in sequential order. The program is operable to index the file using the first and second markers. This index can then be used to provide quick access to individual images stored within the single file.
Preferred exemplary embodiments of the invention will hereinafter be described in conjunction with the appended drawings, wherein like designations' denote like elements, and wherein:
In accordance with the invention, client computer 22 includes a computer readable memory, such as a fixed hard drive shown at 40 containing magnetic media for digital storage of a user interface client application 42 that includes a user interface program along with a number of additional software components. The user interface program is in computer readable form (such as an .exe file). The additional software components include Microsoft™ Internet Explorer™ Internet extensions (included with Windows™ 95, 98, ME, 2000 and revisions), the Win32API libraries (included with Windows™ operating systems), the WinInet.dll (also included with Windows™ operating systems), and a compression library such as IJL115.dll (available from Intel™) to decode and recompress jpeg images received from the cameras 24.
Although the user interface client program 42 is shown as being stored on a hard drive 40, it will be appreciated that it can be stored on other digital storage devices as well. As used in the specification and claims, “digital storage device” includes any of a variety of different digital storage devices, including magnetic media such as a hard disk or removable disk, optical storage media such as a CDROM or DVD, or magneto-optical media.
In general, the user interface client program 42 is operable to access locally stored camera data that uniquely identifies the cameras 24 and then attempts access to those cameras over the network 28. The program 24 is operable to verify access to at least those cameras 24 that are currently accessible, and to generate a user interface display 44 (on the computer's monitor) that includes a display window 46 for each of the cameras 24 accessed over the network 28, and to display in each of the display windows 46 an image 48 received from the camera associated with that display window. As used in the specification and claims in reference to the program 42 or other executable program code, the phrase “operable to” [carry out some action] means that, when executed by a microprocessor or other processing device of the computer, the program directs the microprocessor to carry out or otherwise cause execution of the specified action.
In addition to the user interface client application 42, the hard drive 40 also contains a stream recorder client application 50 and an image viewer 52. Stream recorder client 50 permits high speed recording of streamed images (video) in a manner that minimizes network bandwidth utilization. Image viewer 52 permits browsing and viewing of archived images and video using a playback screen display 54. These two programs 50, 52 are discussed farther below in connection with
The user interface client application 42 is depicted in various levels of detail by
Thereafter, program 42 generates the user interface display 44 and the user can start accessing images or video or can access various menu options shown in
For each camera in the database 56, all control information assigned to that camera is stored in an array data type. This control information includes: Server Enabled, Server Name, Server Model, Server ID, Server IP, Server Mac, Server User ID, Server PW, Server Port, Cam ID, Cam Model, Cam Enabled, Cam Name, Cam Location, Cam Bldg, Cam Room, Cam Tel, Cam Room Contact, Cam Room Contact Tel, Cam Notes, Board Switch Port Nums, Cam Thumb Image Name, Cam Fullimage Name, Record Image Name, Cam Thumb Compression Level, Cam Full Compression Level, Cam Record Compression Level, Tool Tip Text, Last Pan, Last Tilt, Last Zoom, Cam PTZ (y/n), Server Switch Action, Switch Caption, Switch Notes, Email Notification, Email Address, Email Message, Enable Audio Alert, Play Audio File, Last Update Time, Time Lapse Interval, Tie Breaker Count, User Disabled, System Disabled, System Disabled Message, Recompression Value, Pan String, Tilt String, Zoom String, Home String. These parameters all control the camera and/or server used by the application and can be changed via a hardware setup form that is provided by the software for the user and that is described further below.
Once all tests are complete, the process moves to
Once the process begins listening for triggers it then finishes configuring the user interface display grid 44 on the client computer 22. The display grid 44 shown in
Once the screen display 44 has been configured, the display loop begins in which the program accesses and displays images from the cameras 24 on the user's screen, with the software periodically updating the display of each image 48. Starting with the first camera displayed, the program sequentially accesses each camera 24 and displays the received image 48 in the box 46 associated with that camera, scaling the image to fit the user's selected view size. The process runs through the display loop once for each camera 24, incrementing the current frame (display box) for each iteration until it has retrieved and displayed a snapshot image from each camera. It then continues looping through the display loop, starting again at the first camera, and runs through the process continuously, sequentially polling each camera and updating each window 46 with an updated image 48 received from the associated camera. As each display window is being updated, it can be highlighted using, for example, a colored border to distinguish it from the remaining display windows. This allows the user to see the sequential polling process and is especially useful where the images change little if at all from one update to the next.
The display loop of
As a part of the display loop, program 42 requests images or video streams from one of the cameras 24. The requests are formatted as standard CGI based URL syntax, i.e.:
The images are downloaded from the cameras 38 and servers 26 to the client computer 22 where they are stored in a memory buffer, which can be implemented in RAM. The buffer is parsed and augmented with formatted date and time stamps, accurate to the millisecond. The image is assigned a name generated by the program based on the date and time of creation. The program 42 then saves the image to the hard drive 40 using the following directory structure:
This directory structure is generated in real-time as the recording executes. The images are saved in one of the following styles, based on user setup preferences:
Before storing the jpeg image files or concatenating the jpegs onto a jpx stream, the images can be compressed using the compression .dll library noted above. In practice, it has been found that good quality recorded images consume an average of 7,500 bytes per image. Higher compression levels can yield a smaller file (down to 2,500 byes) at a reduced quality. Higher quality images can also be recorded at a file size of 70,000 bytes per image up to 250,000 bytes per image.
The program 42 can acquire images from the cameras in any of three modes:
Snapshot: Request one image from the server and close the connection. By running through the display loop rather than requesting a continuous stream from the server, the program iteratively requests snapshots from the cameras and servers thereby providing continuously-updated images in a manner that requires little bandwidth, with network utilization typically peaking at no more than six percent, regardless of network size.
Stream Frame Mode: Connect in stream mode to each camera and sequentially cycle through each of the cameras, obtaining one or more of the streamed images from one selected camera, pausing the streaming from that camera, and then obtaining one or more streamed images from the next camera until all of the cameras have been accessed and then paused. This cycle is repeated continuously until the user switches the program back out of this stream frame mode. In this mode, the processing is identical to that of the snapshot mode, with the display loop extracting at least one frame from an open stream, and then indexing to the next camera. The connection is periodically reset to remain robustness. This approach to image acquisition yields substantial performance gains over the snapshot mode and is more efficient that a continuous stream. This connection allows the video server to maintain an authorized connection to the client computer, providing enhanced performance and ultimately, augmented received frame rates. Running in this mode requires more bandwidth than the snapshot mode, but will utilize less than the full streaming mode described below, since each stream is paused as the remaining connected cameras are sequentially accessed for a frame of video. The video server buffers surplus video to the point of filling up its internal buffers, at which time the server stops grabbing video frames until the buffer is depleted, or the server is reset. Thus, as will be appreciated by those skilled in the art, this approach can also be used to control the frame rate of displayed and/or recorded video in a manner that minimizes the network bandwidth utilization. Thus, it can be used for only select ones of the cameras and for recording at a frame rate faster than that obtained by individual URL calls, but less than that provided by full streaming from the camera or video server.
Full Streaming Mode: Launch the stream recorder client 50 as a shelled process, pass the URL and various options via a command line syntax. The client computer connects in stream mode and retrieve images in an infinite loop. The connection is reset periodically to ensure robust connectivity across various network topologies. The received image stream is then parsed for SOI (ASCII 255+ASCII 216) and EOI (ASCII 255+ASCII 217) markers, which are the standard start-of-image and end-of-image markers. A buffer is built until these markers are found. When found, the image is displayed and that section is removed from the buffer. This process is then reiterated until a timer expires the session or the user manually closes the connection.
When recording, the stream recorder client 50 writes directly to the same directory used by program 42, but uses a different naming convention to avoid contention. This naming convention can be implemented by appending a letter suffix to the filename of the jpx stream. For example, program 42 might record a filename as “11 00 00.jpx”, whereas program 50 might record images obtained at the same time as “11 00 00_s.jpx”. Since they are now uniquely named, there is no concern of one program overwriting the image files of the other.
Referring now to
The HighliteFrameColor specifies the hex RGB value for the current update window, that is, it identifies the color used for the border placed around the camera window currently being updated at any one time during the display loop of
The RecordMode setting can either be SINGLE or MULTIPLE, which respectively correspond to the single image jpeg and video stream jpx file formats discussed above. For “Single”, the received images are written to disk (or other non-volatile storage media) as a single jpeg file using the naming convention provided above. For “Multiple”, the images are written as jpx streams, with the RecordMode parameter (“24” shown in
The RecordThumbnails=TRUE setting is used to tell the software to record the displayed thumbnails rather than recording the image at the Record Image parameter that is specified in the Camera Hardware Setup form of
The RecordlmageBasePath sets the base directory path for recording all images; that is, it sets the “mapped drive and root path” shown at the upper level of the directory structure shown above. The program will create all lower levels of the directory structure if they do not already exist. An error message is displayed if the drive is not available or the directories cannot be created or accessed, and no recording will occur.
The Camera_Database setting is a pointer to the database 56 that is used for all camera and camera server settings. Other settings that may be less often accessed are:
DisplayWidth: Sets the default user interface display 44 size in twips.
BorderWidth: Sets the width of the black spacing between all thumbnail images in pixels.
SkipInactivePorts: TRUE—Thumbnails that are disabled will not be displayed when program 42 is running.
ShowHighliteFrames: TRUE to show border around currently updated thumbnail.
ShowRecordindicators: TRUE to show the record LED status indicator that can be included at a corner of the display window for each camera.
ShowCameraCaptions: TRUE to show the title caption at the bottom of the thumbnails.
ShowPortNumbers: TRUE to show the physical thumbnail port numbers.
PingTimeout: The number of milliseconds to attempt a ping to the video server before failure abort.
PingOnImageRequest: TRUE to allow ping tests on each received image burst.
CameraCaptionForeColor: Color of the camera caption in hex RGB
CameraCaptionBackColor: Back color of the camera caption in hex RGB
HTTPPort: Http port number to access all cameras. This is a global setting.
MaxFailOverCount: Number of failed camera requests before shutting down the camera port. The display window 46 will show “Unavailable”.
ReconnectAttemptlntervalSecs: Seconds to elapse before trying to reconnect a failed camera.
DefaultEmailRecipient: Email address to use when sending a motion alert.
EmailUserID: UserlD used for email account access.
EmailUSerPW: User password for email account access.
ShowMotionBorder: TRUE to allow the video motion detection routine to highlight the target window when motion is detected.
EnableStatusLog: TRUE to allow status info to be written to a text log.
EnableErrorLog: TRUE to allow specific Error information to be written to a text log.
StreamViewMode: When launching a streamer, the mode it starts in can be:
StreamIntervalSeconds: Number of seconds to allow the streamer to run before closing:
StreamMaxFPS: Maximum number of FPS to stream; 0=Max speed
StreamRecord: 0=no record on program launch, 1=record on launch.
Referring now to
For the server hardware setup tabs of
If the Server Root ID and Server Root Password are not used, the Host Scripting tab will not function. The Connection panel also includes two buttons:
The Switch Inputs tab contains the information related to the digital inputs on the server and can be used to set the program's preference settings for trigger responses from the server. This includes the Show Motion form, Flash Window, and Record Image options shown in
The Relay Output tab provides a display that permits editing of functional settings for latching or pulsing the server relay output. As is known to those skilled in the art, this relay can be attached to a siren, autodialer, or other device compatible with the characteristics and limitations of the server's relay. The fields are:
The Host Scripting tab provides a display for programming or displaying basic CRON scripts on the server. The basic script allows the server to broadcast trigger events from the switch inputs. That is, the script enables to server to send messages to program 42 that are specific to triggering events on the server such as switch closures, infrared, microwave, magnetic, or other forms of sensors. For example, the base CRON script used in Axis™ 2400 series servers is:
where xx is the server port 1-4, yyy.yyy.yyy.yyy is the IP address of the client computer 22, and zzzz is the port specified in the .ini file for communication between the program 42 and the server. The fields on this tab are:
The Host Scripting panel also contains four buttons:
For the camera hardware setup tabs of
The Presets tab allows the user to define any quantity of user defined presets for Pan/Tilt/Zoom. The user is shown a thumbnail of a static snapshot for any new positional requests. The fields are:
There is one button:
There are three slider controls:
The Displayed Image tab allows the user to set the image size and compression level request for the camera server image retrieval. These parameters are mapped against CGI type parameters defined by the server. The user representations are simplified terms versus the CGI syntax. Three image styles are defined here. Each style is set for the three image request types. This value is used by the server when sending requested images to the application. The fields are:
The Thumbnail Image is displayed on the Main form with all the other camera views. The Full View Image is displayed on the Motion form. The Record Image is used whenever a save request is made by the application. Each Image style has a related compression value. The compression values are selectable for each Image style. The following considerations should be used when selecting the Image styles:
The Switch Inputs tab allows the user to bind any or all available digital switch ports to a display window. That is, the digital switch inputs must be bound to camera ports. This setting tells the application which cameras to record/notify when a trigger event occurs and allows the application to notify the user on screen which port trigger(s) is/are associated with a camera view. Each defined switch input is listed with a check box that can be selected to bind the camera to that switch input.
As mentioned above in connection with
The compression value is used by the compression library to set the amount of compression for image saves. The lower the number, the higher the compression. The range is between 5 and 40. The most commonly-used value is between 30 and 35.
The Email tab permits the user to set an email address to which an email notification will be sent upon occurrence of a trigger event. An image from the triggered camera can be attached to the email. The Performance tab permits the user to specify a unique User ID so that the camera or video server administrator can set or limit the bandwidth of video transmission on a user-by-user basis.
When a user double-clicks on a camera window 46, the program 42 brings up a Motion Form which runs the process shown in
In installations where very high speed recording is being performed, the jpx file can become quite large. A typical jpx file for an hour of video is 1 to 5 MB, while high speed recording can easily exceed 20 MB and even reach 100 MB. Since the jpx is purely concatenated JPG images, aligned head to tail, the only way to view the images within the jpx file is to parse the SOI and EOI markers and display the image in between them. For incremental playback, this process is easily controlled. However, when a user wants to “rewind”, “fast forward”, or randomly jump to a portion of video, this process requires the program 52 to process from its current position and read all video frames in between. Jumping over frames does not work for these files since, in order to most efficiently store the images, the images within the jpx file are relatively unique in length. This process of reading in all of the frames can be undesirably slow when the jpx file is large. To overcome this issue, the image viewer program 52 creates an index of the SOI and EOI markers of each image in the jpx file. This index is stored as an array of pointers into the jpx file and is stored using .ndx as its filename extension. The program 52 will look for this index and use it to load the image pointers instead of reading the entire jpx. The image pointers identify the memory locations of the SOI and EOI markers, thereby allowing the program to easily locate and retrieve individual images contained within the jpx file. The ndx file can be loaded in less than a second, as compared with jpx load times of up to a minute or more, depending on jpx size. The index is created on the first user access to the jpx image file or when an autoindex event is processed, as described above. Successive requests from the user to a jpx will allow the program to read the ndx file rather than a conventional approach which would entail reading the entire jpx file. With the index file, the pointers can be read into memory and accessed quickly, even for very large jpx files.
The ndx index file also makes possible the use of the image viewer program 52 as a plug-in for commercially available web browsers. In conventional browser-based retrieval of video, the video files can be quite large, and the user can therefore typically only request a small number of frames. The user does not know the stream size, frame count, index of time pointers, etc. By implementing the image viewer 52 as a plug-in component for a web browser, viewer can be used to provide web page access from a web browser to stored jpx images. This alleviates the necessity of a specialized software product to view the archives. Since this image viewer plug-in utilizes the ndx index file, the user can retrieve a wealth of video specific content in a very short period of time, allowing realtime display of video (assuming the user has sufficient bandwidth) of the Internet or other network.
When implemented as a plug-in, the image viewer 52 can also be used to view “live” a stream that is being recorded on a network server. This can be accomplished using server software that records the jpx files in the manner described herein, with the image viewer then only parse the target jpx file on the network server for the latest video frame. By displaying this video frame, the user sees live video that originated from the camera server, but is being supplied by what can be a much higher performance network server. Thus, since the user is now viewing recorded video from a higher performance server, the number of simultaneous connections to the video feed is highly scalable by implementing standard web server components. This is advantageous since all Ethernet video servers are highly susceptible to overload from multiple user requests and performance is degraded linearly with each simultaneous connection.
Rather than using the ActiveX control provided by the camera/server manufacturer, the user interface client program 42 and stream recorder program 50 utilize their own ActiveX control to interface with the cameras and servers. An overview of this OCX control is shown in
Referring now to
Referring now to
The process begins by taking the first image (i.e., frame) of a video sequence and placing it into a buffer where it will be used as a reference image. The reference image is compared to subsequent video frames on a pixel-by-pixel basis. Preferably, the reference image is compared to subsequent images as they are being received (i.e., as soon as they are received) by the client computer so that the program is operable to provide real time motion detection. For each pixel in the current image, the program performs a comparison of the color component values for that pixel with the color component values of the corresponding pixel in the reference image. If the difference in color component values for corresponding pixels from the two images differ by more than a preselected amount, the program generates a motion detect signal which can be implemented as the Record Flag shown in
Once motion is detected, the system records video and continues the pixel comparisons for each subsequent frame until the routine detects that motion is no longer present. To prevent the recording from ending before a subject has completely left the camera's field of view, the program continues recording until a specified number of extra frames have been recorded. This is carried out using an Extra Frames counter that is decremented once per frame starting with the first frame received after no further motion has been detected. As with the other counters, the number used for this counter can be user specified. Periodically, a new reference frame is selected from the video stream. This is done at regular intervals as specified by a user “Refresh Interval” setting, which is an image frame count. This Refresh Interval is used in conjunction with a frame counter C_Ref such that, once the C_Ref counter exceeds the Refresh Interval number, the reference buffer is emptied and the counter reset to zero.
The core function of the video motion detection is based on a count of pixels for which at least one color component value differs from that of the reference image by a preselected amount. Minimum object size detection can be implemented using the routine of
Apart from minimum object size detection, the program 42 is also operable to permit the user to specify a region of the camera's field of view so that the program performs the pixel comparisons only for those pixels located within that region. This region of interest processing is implemented by providing the user with the ability to mask out specific portions of a video frame, which will then be ignored by the motion detection. This concentrates the motion detection to specific regions of the screen. Typical region of interest masks employ grids or regions based on a grid array of squares. These areas are selected to mask out the regions. This tends to be somewhat granular in use. The program 42 uses a pixel based approach to masking the images. The user first creates a region of interest by defining the mask using the computer mouse to paint a black or other colored area on the image. The painted areas represent those portions not desired for use in detecting motion. A tool palette of square or other shape may be employed to create the painted mask. The mask is converted to a black and white mask and saved as a .bmp file based on the camera name. When video motion detection is used by the program, individual pixels from the reference and current images will be compared only if the corresponding mask pixel is not black. Users may edit the mask in Microsoft™ Paint or other image editor to refine the mask to the pixel level.
Implementations of this type of streaming media have a common difficulty; namely, Ethernet connectivity is at best a highly reliable but not a totally reliable connection. Occasionally, power fluctuations, excessive image requests, overloaded or failing networks and a multitude of other issues can cause the connection to the camera to fail. Competing implementations often cause the application to abnormally abort or stop responding, requiring the user to close and restart the application. This condition is very serious since recorded video will not be available while the program is not responding. The user interface client program 42 and stream recorder client program 50 address this issue by identifying three modes of failure, and providing contingency functions to overcome these problems. These modes are:
Failure types 1 and 2 are easily overcome by the implementation of the default timeout of failed requests by the Wininet.dll and its related components. The program can count these failures on a per camera basis and optionally shut down the camera and reattempt access periodically. This maximum failure count and reattempt connections are user selectable on a global scale.
Type 3 failures are usually serious and can cause the program to stop responding. By executing the request asynchronously or in a worker thread, the request can be timed and if abnormally long, cancel the request and try again. This allows the application to continue to function consistently. Preferably, the program utilizes both methods to insure robust connectivity.
It will thus be apparent that there has been provided in accordance with the present invention a digital video system and computer program therefor which achieves the aims and advantages specified herein. It will of course be understood that the foregoing description is of preferred exemplary embodiments of the invention and that the invention is not limited to the specific embodiments shown. Various changes and modifications will become apparent to those skilled in the art. For example, although the invention has been described as it would be implemented using TCP/IP network cameras and camera servers, it will be appreciated that the invention can be used in conjunction with other image sources. For example, rather than specifying a URL to access a particular camera, the invention could be used to specify a memory location or utilize a .dll file to access images from, for example, a video capture card, USB or IEEE-1394 (Firewire) ports. All such variations and modifications are intended to come within the scope of the appended claims.