TWI601425B - A method for tracing an object by linking video sequences - Google Patents

Description

Method for serially connecting photographic pictures to form an object trajectory

The present invention relates to a multi-camera monitoring system, and more particularly to an object serial connection correction method for correcting an object of a multi-camera monitoring system that erroneously cascades multiple photographic pictures, and a multi-camera monitoring system using the photographic picture serial connection correction method.

The traditional camera monitoring system provides a specific event detection service for a single monitoring area, and returns all relevant video data and detection results to the central server. However, in the application of video surveillance, the provision of specific event detection services for a single monitoring area has been unable to meet the needs; especially for post-analysis applications, it is often necessary to carry out relevant human activities throughout the monitoring system for the events that have occurred. When a complete description of the temporal location trajectory occurs, it is not possible to provide a specific event detection service for a single specific environment. Accordingly, multi-camera monitoring systems have become the mainstream in today's surveillance systems.

In the multi-camera monitoring system proposed today, each system transmits a photographed image taken by each camera set in a specific monitoring area to a central server, and the central server performs image analysis on the content of the images taken by each camera. To obtain the result of object analysis in a single screen. Then, the central server obtains the temporal and spatial correlation of each object in each photographic picture according to the result of the object analysis (that is, the correlation between the order of each object appearing in each monitoring area and the position), and according to the space-time association of each object. Sexually connect specific objects to obtain trajectory information and historical image sequences of specific objects in the overall multi-camera monitoring environment.

Please refer to U.S. Patent No. 7,224,423, the name of which is "Linking zones for object tracking and camera handoff". The multi-camera monitoring system provided by this related patent independently performs image analysis on the video data captured by each camera, thereby obtaining the analysis result of detection and tracking of each object in the monitoring range of a single camera. Then, the camera monitoring system extracts the position and appearance of each object in the monitoring range of each camera according to the analysis result, and then establishes a probability distribution function according to the appearance and departure time of each object and its time correlation. Probability distribution function. In this way, the camera monitoring system can estimate the relationship of the objects appearing on each photographic picture through the above-mentioned probability distribution function, thereby connecting the specific objects in each photographic picture, thereby obtaining the specific object in the overall multi-camera. Monitor historical imagery and trajectory information in the environment.

In addition, please refer to the published patent application of the Republic of China TW200943963, the name of which is "Integrated Image Surveillance System and Method / INTEGRATED IMAGE SURVEILLANCE SYSTEM AND MANUFACTURING METHOD THEREOF". This published patent application proposes an image registration method for splicing a plurality of photographic images taken by a multi-camera into a single picture to reduce the user's monitoring burden. Although the splicing of a plurality of photographic images taken by a plurality of cameras into a single picture can effectively reduce the monitoring burden of the user, the related patent application does not propose a corresponding intelligent multi-camera monitoring content analysis system. Although the spliced multiple single images can be used as the video data of the multi-camera monitoring system, the resulting spliced single image is too large, which still causes the computing burden of the intelligent multi-camera monitoring content analysis system.

The aforementioned multi-camera monitoring system trusts the image analysis and object serialization algorithms used, and automatically concatenates the same specific objects for each photographic image to generate a trajectory image of a specific object. However, because of the actual environment, it will be possible to produce different degrees of error for various algorithms. Therefore, the aforementioned multi-camera monitoring system may incorrectly connect different objects in series, and cannot be corrected immediately.

Embodiments of the present invention provide a series connection of objects of a photographic picture obtained by a multi-camera monitoring system and a method for correcting the same. The object serial connection of this photographic picture and its correction method are used in a multi-camera monitoring system, and the steps thereof are described below. A user interaction platform is provided to the user to allow the user to select a specific item to be tracked through the user interaction platform. According to the current shooting time of the photographic picture as a boundary, before and after this time, the photographic pictures of related objects appearing in the monitoring screen of each camera and related to the specific object are respectively displayed on the user interactive platform according to the time sequence. A list of previously related objects and a list of related objects. In addition, the system scores relevance based on the correlation between objects, which in turn results in the splicing of objects across the camera. According to the foregoing time definition, the serial connection result is respectively displayed in the previous serial connection result of the area object and the subsequent serial connection result of the object, that is, the trajectory image sequence of the specific object to be serially connected appears in At present, several photographic images taken before the shooting time of the photographic screen and taken by other cameras are arranged in the chronological order of the previous objects of the user interaction platform according to the chronological order and the relevance score. The trajectory image sequence of the specific object that will be automatically connected in series, which appears after the shooting time of the current photographic image and is taken by other cameras, and the subsequent objects arranged in the user interaction platform according to the chronological order and the relevance score. Concatenated results. The user determines whether the concatenation result is correct by referring to the previous related object list, the subsequent related object list, the previous object concatenation result, and the subsequent object concatenation result. If the result is found to be incorrect, the relevance rating in the object list is clicked. Not the largest specific item in a particular photographic picture, thereby instructing the multi-camera monitoring system to correct the automatic splicing result of a particular object.

The embodiment of the invention provides a multi-camera monitoring system, which comprises a plurality of video capture and analysis units, a plurality of video analysis data collection units, a video analysis data database, a multi-video content analysis unit and a user interaction platform. Each of the video capture analysis units is implemented by a camera connected to a video analysis device and configured at various locations in a monitoring environment of the multi-camera monitoring system. The video analysis device can be a computer or an embedded system. Each video capture unit is connected in series with the associated video analysis data collection unit. The video analysis data database is connected in series with a plurality of video analysis data collection units, and the multi-video content analysis unit is connected in series with the video analysis data database. The user interaction platform is connected to the video analysis data analysis unit, which is used for the user to select a specific object to be tracked, and allows the user to refer to the user related platform to provide a list of previously related objects, a list of subsequent related objects, and previous objects. The result of the concatenation of the concatenation result and the subsequent object is selected by the selected correction object in a specific photographic picture whose relevance in the subsequent object list is not the largest, thereby instructing the analysis unit to correct the automatic concatenation result of the specific object.

The user interaction platform is based on the shooting time of the current photographic screen, and before and after this time, the photographic images of related objects appearing in the monitoring screen of each camera and related to the specific object are respectively displayed in time order. A list of previously related objects of the interactive platform and a list of subsequent related objects. In addition, the system scores relevance based on the correlation between objects, which in turn results in the splicing of objects across the camera. According to the foregoing time definition, the serial connection result is respectively displayed in the previous serial connection result of the area object and the subsequent serial connection result of the object, that is, the trajectory image sequence of the specific object to be serially connected appears in At present, a plurality of photographic images including the specific object captured by other cameras before the shooting time of the photographic screen are arranged in the chronological order of the previous objects of the user interaction platform according to the chronological order and the relevance score. The trajectory image sequence of the specific object to be automatically connected in series, which appears after the shooting time of the current photographic image, and which is taken by other cameras, includes several photographic images of the specific object, and is arranged in time series and relevance scores. The subsequent objects of the interactive platform are concatenated in the results. The user selects a specific object in a specific photographic picture whose relevance score in the subsequent object list is not the largest by referring to the previous related object list, the subsequent related object list, the previous object splicing result, and the subsequent object splicing result. Instructs the multi-camera monitoring system to correct the automatic concatenation of specific objects.

In summary, the multi-camera monitoring system provided by the embodiment of the present invention has a method for correcting the object serial connection of the photographic image, and the multi-camera monitoring system has a user interaction platform for the user to operate, and the user performs the photographic image. The object serial connection correction method is to correct the error that may occur in the automatic serial connection object of the conventional multi-camera monitoring system.

The above and other objects, features and advantages of the present invention will become more <RTIgt;

In order to fully understand the present invention, the embodiments will be described in detail below with reference to the accompanying drawings. However, it is to be noted that the following examples are not intended to limit the invention.

Please refer to FIG. 1. FIG. 1 is a block diagram of a multi-camera security monitoring system according to an embodiment of the present invention. The multi-camera monitoring system 100 includes a plurality of video capture analyzing units 110, a plurality of video analytics data collecting units 120, a video and analytics data database 130, a multi-video content analyzing unit 140, and a user interaction platform 150, wherein the plurality of video cameras 多个The fetching analysis unit 110 is placed at a plurality of different locations to monitor for a plurality of different monitoring zones. Each of the video analysis and analysis units 110 is connected in series with a corresponding video analysis data collection unit 120, and the video analysis data collection unit 120 is further connected to the video and analysis data database 130. The video analytics data repository 130 is coupled to the multi-video content analysis unit 140, and the multi-video content analysis unit 140 is coupled to the user interaction platform 150.

The video capture analysis unit 110 is configured to obtain a photographic image of the monitored monitoring area, and perform image analysis on the photographic image, thereby extracting the physical information of each object and each object in the photographic image to obtain object analysis. result. Then, the video capture and analysis unit 110 transmits the image sequence and the object analysis result to the corresponding video analysis data collecting unit 120. The captured image captured by the video capturing and analyzing unit 110 at a plurality of consecutive time points may constitute an image sequence.

In more detail, the video capture analysis unit 110 can be implemented by a digital camera serially connected to the video analysis device. The video analysis device can be implemented by a computer or an embedded system platform. The digital camera is used to capture the photographic images at each time point, and the video analysing device analyzes the acquired photographic images, and obtains the analysis results of the objects, such as the number, position and features, which are unique to the analysis, and then analyzes the results of the objects. And the image sequence is transmitted to the video analysis data collecting unit 120.

In order to efficiently transmit the image sequence and the object analysis result, the video analysis data collecting unit 120 is configured to compress and edit the received object analysis result and the image sequence to generate a data compression editing result. Then, the video analysis data collecting unit 120 stores the data compression editing result into the video and analysis data database 130, wherein the data compression editing result has the object analysis result and the photography picture information.

In more detail, the video analysis data collecting unit 120 performs data compression editing on the image sequence through a video compression method (such as a video encoding method with high compression effect such as H.264) to reduce the transmission bandwidth requirement. In addition, for the object analysis result, the video analysis data collecting unit 120 first inserts the time series information into the object analysis result, thereby confirming the correspondence between the object analysis result and the photographing screen. Then, the video analysis data collecting unit 120 performs corresponding information conversion (such as data compression) according to the use, so as to reduce the amount of data transmitted.

In order to effectively synchronize the relationship between the photographic image and the object analysis result, and reduce the amount of data transmitted, the video analytics data collecting unit 120 can insert the time series information into the object analysis result, and the video analytics data collecting unit 120 can also use the data hiding technology or use the video. The user data zone defined by the compression standard hides the object analysis result of each photographic image in the video material corresponding to the image sequence. For example, the video analysis data collecting unit 120 sequentially hides the bit data of the processed image analysis data from the bit data of the processed image analysis data in the data structure corresponding to the image sequence through the data hiding method. In the discrete cosine transform (DCT) parameter.

The video and analysis data database 130 is used to store the data compression editing result transmitted by the video analysis data collecting unit 120. Since the data compression editing result has the object analysis result and the information of the photographing screen, the video capturing and analyzing unit 110 is in the monitoring area. The temporal and spatial relationship between the captured photographic screen and the appearance of the object is stored in the video and analysis data library 130, so that the multi-video content analyzing unit 140 reads the data required for analyzing the specific object.

The multi-video content analyzing unit 140 reads the data required for analyzing the specific object from the video and analysis data library 130, and analyzes the correlation between the objects in the respective photographic images of the video capturing and analyzing unit 110 and the specific objects. To serialize the complete historical trajectory information of the specific object currently analyzed, and then generate a historical image sequence of the specific object.

In more detail, the multi-view content analysis unit 140 reads the data required for analyzing the specific object from the video and analysis data library 130, and extracts the corresponding object analysis result embedded in the data, thereby analyzing the specific Correlation scores are obtained between the objects and the objects in the respective photographic images, and the correlation analysis results are obtained. Then, the multi-video content analyzing unit 140 concatenates the specific objects appearing in each photographic picture according to the correlation analysis result, and generates a concatenation result with the largest relevance score, wherein the concatenation result with the largest relevance score is specific. The trajectory image of the object. Then, the multi-view content analysis unit 140 provides the generated trajectory image of the specific object to the user through the user interaction platform 150, and feeds back the video data corresponding to the trajectory image of the specific object to the video and analysis data database 130. Store.

In other words, the specific object across the monitoring area of the video capture analysis unit 110 is recorded in the photographic image of each video capture analysis unit 110, and the multi-video content analysis unit 140 can display multiple photographic pictures according to the chronological order. The specific objects are connected in series to form a trajectory image of the specific object. The trajectory image of the specific object allows the user to quickly view when the specific object appears and leaves in each monitoring area through the user platform 150, thereby learning the complete behavior history information of the specific object in the overall multi-camera monitoring environment.

The user interaction platform 150 can provide the user to obtain the photographic images of each monitoring area from the video and analysis data database 130 and perform synchronous playback control directly for each monitoring area. In addition, the user interaction platform 150 can also perform specific event detection and specific object tracking according to the monitoring conditions set by the user.

In addition, the user interaction platform 150 of this embodiment can also provide the user with a specific object, in view of the fact that the conventional multi-camera monitoring system cannot guarantee that the trajectory image of the specific object that is automatically generated is the correct cascading result. The trajectory image is corrected so that the last rendered trajectory image is the correct splicing result.

Corresponding to the user interaction platform 150 having the ability to correct the trajectory image, the multi-video content analysis unit 140 must provide a string with other relevance scores in addition to the cascading result with the highest relevance score. The result is obtained, so that the user directly corrects the trajectory image of the specific object on the user interaction platform 150, wherein the serial connection results are arranged according to the level of the relevance score.

If the user does not correct the trajectory image through the user interaction platform 150, the multi-video content analysis unit 140 defaults the cascading result with the highest relevance score to the correct splicing result, and then continues the subsequent splicing work to Produces a trajectory image of a specific object. Conversely, if the user considers that the concatenation result of the highest relevance score generated by the multi-video content analysis unit 140 is an incorrect concatenation result, the user can select other concatenation results through the user interaction platform 150 to borrow This corrects the concatenation error and produces the correct trajectory image for the particular object.

For example, the multi-video content analysis unit 140 can access the video and analysis data database 130 to view the edited data of the desired analysis according to the time of the event scheduled by the user on the user interaction platform 150. Next, the multi-view content analysis unit 140 analyzes the objects dispersed in the respective photographic images, such as the date, time, weather, and the like of the appearance, departure, trajectory, object features, and even the appearance of the objects in the past history of the objects. By analyzing the information, the multi-video content analyzing unit 140 can understand the probability values of the various objects in various conditions in the monitoring environment and the trajectory distribution that may travel, and can obtain the correlation analysis result of each object, and thereby serially connect the objects. The same objects scattered in the photographic images taken by the different video capture analysis units 110 are obtained, thereby obtaining a complete trajectory of all the objects in the overall monitoring environment.

The object analysis result obtained by the video capture analyzing unit 110 can be linked with the correlation analysis result obtained by the multi-video content analyzing unit 140. In order to make each object appear in the corresponding photographic picture, the object information can be displayed in the monitoring environment in which the object appears in the monitoring environment in which the object number, the position and the extracted object feature are displayed, and the multi-video content analyzing unit 140 will have multiple possible The information such as the number of the object, the probability of occurrence, and the location of the monitoring environment are aggregated into correlation analysis results, and the correlation analysis result is embedded in the corresponding video data. Then, the multi-video content analyzing unit 140 feeds back the video data embedded in the correlation analysis result to the video and analysis data database 130 for storage, so that the user interaction platform 150 presents the required content.

The multi-view content analysis unit 140 can use the object information of the previous and current space, time, object features, and the like of the object to concatenate the same object of each photographic picture. The above object information can be divided into three levels according to the difficulty in obtaining the analysis data, the order in which it appears, and its characteristics.

The object information of the first level is the position and speed of the object. By the position where the object appears, disappears, and the speed of travel at that time, the multi-video content analyzing unit 140 can estimate the position at which the object may appear next time. In more detail, the multi-view content analysis unit 140 can cooperate with the projection theory by using the information such as the appearance and disappearance of each object in each photographic image and the spatial position of the video information of the analysis unit 110 set by the user. Then construct the probability distribution function (PDF) of each object. Next, the multi-view content analysis unit 140 performs the relevance score using the probability function, thereby concatenating the same objects distributed in the respective photographic pictures.

For example, in the monitoring environment of the MRT station, for a person who has just entered the entrance of the MRT station, the highest probability of the position that will appear next time should be the monitoring area near the entrance of the MRT, such as the toll gate. position. In contrast, since the object has not passed the toll gate, the probability of the object appearing on the waiting platform is zero. Accordingly, the multi-video content analyzing unit 140 analyzes the probability distribution function of the person in the monitoring area where the video capturing and analyzing unit 110 is located next time when the person appears in a certain position. In other words, the multi-view content analysis unit 140 can use the probability distribution function to serially connect the same objects appearing on each monitoring screen to obtain the trajectory information and the historical image of the same object.

The object information of the second level is the object feature, and the multi-video content analyzing unit 140 can be used to compare the objects appearing in the monitoring area at different times, and according to the same object appearing in each photographic picture. In more detail, the multi-view content analysis unit 140 obtains possible candidate objects of the object to be concatenated according to the probability distribution function, and filters out the lower possible possible candidate objects by using the analyzed object traveling direction. Then, the multi-video content analyzing unit 140 converges the objects by comparing the object features (such as color, shape, and the like) of the object in the photographic image, that is, when performing correlation analysis, considering the probability distribution function and the object feature information. And then obtain a better correlation score result.

Taking the monitoring environment of the MRT station as an example, the multi-video content analyzing unit 140 analyzes the speed and direction of the departure gate gate and the corresponding position of the monitoring image of the personnel leaving the toll gate by the person leaving the toll gate of the MRT station. Then, the multi-view content analysis unit 140 analyzes the photographic images of the video capture analysis unit 110 that may appear in the next moment, and compares the characteristics of the people in the photographic images of the video capture analysis unit 110. (such as color) and the behavior track of the people in each photographic picture.

The object information of the third level is historical data, and the multi-video content analyzing unit 140 can count the past video data, and then analyze all possible moving trajectories of each object, calculate the probability of distribution of various trajectories, and estimate the possibility of analyzing the objects. The location appears. In more detail, the multi-view content analysis unit 140 can perform data analysis and data statistics on all historical data (previous video data) in the monitoring environment and the object information analyzed therein to obtain a corresponding monitoring environment. Relatively credible object statistics. In addition, the object statistical information can be further classified according to different conditions such as time and environmental parameters. In this way, the multi-view content analysis unit 140 can know the historical behavior track of the object in the environment under the condition of specific time and environment parameters. That is to say, when the correlation analysis is carried out, the probability distribution function, the object feature information and the historical object behavior trajectory classification information are simultaneously considered, and the better correlation score result is obtained.

Taking the monitoring environment of the MRT station as an embodiment, the multi-video content analyzing unit 140 analyzes and compares the past video data of a MRT station, and after counting the video sequence of a certain length of time, it is known that the personnel are in the class time. Historical behavioral trajectory. For example, during the class time, the people wearing the student uniforms will only pass through the entrances and exits, cross the upper MRT passage, and then leave the MRT station, and will not enter the MRT station for the MRT; Most of the people entering the MRT station will pass through the MRT station entrance and exit, cross the toll gate, and then take the MRT.

Thus, the multi-video content analyzing unit 140 can know the flow statistics of the inbound and out of the MRT station, and thereby estimate the possible traveling direction of the person appearing in the monitoring area. For example, in the time of class, if a person wears a specific student uniform, the probability that the person will leave the MRT station through the upper channel will be higher than the probability of taking the MRT through the toll gate.

In addition, when the objects of the photographic images of the video capture analysis unit 110 are concatenated, the multi-view content analysis unit 140 scores the correlation of the object trajectory distributions. Each node in the relevance score extension map is represented as a possible object for tracking. By tying up all possible objects with the highest relevance score, the behavior track of the object will be obtained.

2 is a schematic diagram of an interface of a method for correcting object serial connection of a photographic image on a user interaction platform according to an embodiment of the present invention. The interface on the user interaction platform includes a monitoring environment window 210, a camera list window 220, at least one monitoring object window 230, and a multi-camera photographic screen window 240. The object serial connection correction method of the photographic image can be implemented by using software, and the interface on the user interaction platform can be realized on the platform of various operating systems. However, the implementation method of the object serial connection correction method and the user interaction platform on the photographic screen is not limited thereto.

The monitoring environment window 210 includes an environment diagram 211 for presenting an overall monitoring environment. The environment diagram 211 can be used to understand the geographical characteristics of the monitoring environment (such as information on corridor location and room layout) and the distribution of the video capture analysis unit 110 (also That is, the distribution location) and the behavioral trajectory of the specific object in the monitoring environment. The user can set the environment diagram 211 to select one of the geographic environment map, the architectural architecture map and the monitoring facility map as the environment diagram 211, or the user may select some or all of the above diagrams to be superimposed and stacked. The combined diagram is taken as an environmental diagram 211. In addition, the user can also present the schematic 211 through 3D computer graphics.

The monitoring environment window 210 also includes a playback control unit 212 and a timeline control component 213. The play control unit 212 is configured to effectively control the playback (forward and backward) of the video data when tracking, presenting and correcting the after-travel historical trajectory of the specific object, and the time axis control component 213 can control the video data to start playing at a specific time point. . The playback control unit 212 can control the playback of all the video data presented in the user interface to synchronously capture the video data of the video capture analysis unit 110 of the multi-camera monitoring system 100 on the interface of the user interaction platform 150. Play.

The camera list window 220 is used to present the relationship between the number of all cameras in the system (i.e., the cameras used in the video capture analysis unit 110) and the position of the camera in the surveillance environment. Each camera can be displayed and distinguished by a specific recognition method, for example, to distinguish each camera in a different color. The contents of the camera list window 220 are displayed in synchronization with the contents of the monitoring environment window 210. When the user clicks on one of the cameras in the camera list window 220, the selected camera will be displayed in the monitor environment window 210 and the multi-camera list window 220 in bold colors, while the unselected camera will be in a non-obtrusive color. The markers are presented on the monitoring environment window 210 and the multi-camera list window 220.

The display screen 231 of the monitoring object window 230 is used to present a photographing picture currently taken by the camera selected by the user. The monitored object window 230 can continue to present the selected item even though the selected item has left the monitored area of the originally selected camera. The monitoring object window 230 allows the user to correct the result of the serial connection of the objects (ie, correct the behavior track of the selected object) to correct the object of the multi-video content analyzing unit 140 erroneously concatenating the photographic picture.

In more detail, the user can select, through the monitoring object window 230, the previous and subsequent possible items (presented by the previous object list 232, the subsequent related object list 233), previously and partially present, which are currently being tracked, previously and subsequently. The concatenation result (presented by the previous object concatenation result 234 and the subsequent concatenation result 235), so that the user can monitor the object window 230 to correct the result of the object concatenation, thereby avoiding the multi-video content analysis unit 140. Error connecting the objects of the photographic screen. At the same time, in order to enable the user to clearly understand the complete state of the possible objects, the previous and subsequent possible objects may be presented as an image sequence or a screenshot of the complete object or an object track image generated by the overlay. Image sequence playback refers to the sequence of images recorded by the possible object within the surveillance range of the camera. The complete object screenshot refers to the complete surveillance image captured by the object when it is completely displayed in the camera monitoring range, and the object trajectory image superimposes the image sequence recorded by the possible object in the camera monitoring range through a specific image processing. A single image of the special treatment produced by the method.

The multi-camera photographic screen 240 is used to present an instant photographic image captured by a plurality of cameras selected by the user, or to play historical video data of a plurality of selected cameras recorded in the video and analysis data repository 130. The multi-camera photographing screen window 240 may be composed of a plurality of video playback windows of a specific combination, or may be a photographing image taken by a plurality of cameras by at least one floating window.

When the user monitors the monitoring environment on the user interaction platform 150, the interface on the user interaction platform has a monitoring environment diagram 210, a camera list window 220, and a multi-camera photography screen window 240. The multi-camera photographic screen 240 presents a number of even all of the instant photographic images, and the rendered photographic images are captured by the video and analytics data repository 130. The photographic screen of each camera can be an independent sub-window 241, and the size and position of each independent sub-window 241 can be set by the user. In addition, the photographing screen of each camera may be one of the divided screens 242, and the distribution manner thereof is set by the user.

FIG. 3A is a schematic diagram of an interface on a user interaction platform when a user selects a camera for real-time monitoring according to an embodiment of the present invention. When the user selects one of the cameras from the multi-camera photographic picture window 240, the camera distribution position in the monitoring environment window 210, or the multi-camera list window 220, the specific camera monitoring window 250 is immediately generated while monitoring the environment window 210 and the camera list. Window 220 marks the selected camera in a bold color (such as red), while other unselected cameras are marked in a non-obtrusive color (such as dark gray). In addition, the multi-camera photographing screen window 240 will be reduced to the lower edge of the screen of the interface; or, the multi-camera photographing screen window 240 will be reduced to be located at the edge of the screen of the interface, and the photographing screens of other multi-cameras will be presented with a reduced screen.

The photographed picture currently taken by the selected camera is presented on the display screen 231 of the particular camera monitor window 250. In addition, the previous related object list 232 will present the captured photographic image of the selected camera adjacent to the camera a few seconds ago, and the subsequent related object list 233 will present the photographic image currently captured by the adjacent camera of the selected camera. In addition, since the user has not selected the object to be tracked, the previous object serialization result 234 and the subsequent object serial connection result 235 do not need to present any content, and may be presented in a dark color, or may not appear in a specific camera. Monitor window 250.

For example, when the user clicks on the No. 1 camera, the specific camera monitoring window 250 corresponding to the No. 1 camera is immediately generated. At the same time, the camera No. 1 in the camera list window 220 will be presented in red, and the remaining cameras will be presented in dark gray. The position A in the environment diagram will present a red frame, and the remaining positions (position B to position H) will present a translucent dark gray frame. Since the live monitoring does not require the use of the playback control unit 212 and the timeline control element 213, the playback control unit 212 and the timeline control component 213 are presented in a translucent manner. In addition, the multi-camera photographic screen window 240 will be reduced to the lower edge of the screen of the interface.

FIG. 3B is a detailed schematic diagram of a specific camera monitoring window according to an embodiment of the present invention. As described above, since the user has not clicked on the item to be tracked, the previous item concatenation result 234 and the subsequent item concatenated result 235 do not need to present anything, and may be presented in a dark color.

In addition to presenting the photographic picture currently captured by the selected camera to the display area 231, the multi-camera monitoring system 100 also marks the number of the selected camera on the specific camera monitoring window 250, for example, marking the No. 1 camera. The upper edge of the window 250 is monitored by a particular camera. In addition, the multi-camera monitoring system 100 can also mark the shooting time on the display area 231.

In addition, the multi-camera monitoring system 100 also captures the object information in the photographic image (including the appearance of the object, the object number and the object feature, etc., but not limited thereto), and marks the information of the object on the display. The object of the photographic picture of the area 231. The position of the object will be marked by a box, and the object information of the object will be described around the box (such as the object number with the highest probability value, the object type, the color characteristics, the spatial information currently existing in the monitoring environment, etc.), but Not limited to this).

For example, in FIG. 3B, the position where the character A appears is marked by a box, and the object information is marked near the box, wherein the object number and the color feature of the character A are 123, person and brown, respectively. Similarly, the position where the character B appears will be marked with a box, and the object information is marked near the box, wherein the object number and color characteristics of the character B are 126, person and red/gray, respectively.

The previously related object list 232, in addition to the captured camera taken by the proximity camera of the selected camera a few seconds ago, may also have the shooting time and camera number marked in the previously associated object list 232. Similarly, the subsequent related object list 233 will display the photographing time and the camera number in the subsequent related object list 233 in addition to the photographing picture currently taken by the adjacent camera of the selected camera. In the previous related object list 232 and the subsequent related object list 233, the order of the photographic pictures is arranged by the camera number or the distance from the selected camera.

In the previous related object list 232 and the subsequent related object list 233 of FIG. 3B, the order of the photographic pictures is arranged by the camera number, so that the photographic pictures taken by the cameras 2, 3, 4, and 6 adjacent to the No. 1 camera will be Arranged in order. In FIG. 3B, the shooting time of the photographed picture currently taken by the No. 1 camera displayed in the display area is 12:06:30, so that the photographs taken by the cameras 2, 3, 4, and 6 of the subsequent related object list 233 are photographed. The shooting time of the screen is also 12:06:30. In addition, the shooting time of the camera shots of cameras 2, 3, 4, and 6 of the previous related object list 232 is 12:06:20.

FIG. 4A is a schematic diagram of an interface on a user interaction platform when a user selects a specific object for real-time monitoring according to an embodiment of the present invention. When the user clicks on a particular item, the particular camera monitoring window 250 will become the monitored object window 230. For example, after the user clicks on the object of the item number "123", the interface on the user interaction platform 150 will change, and the specific camera monitoring window 250 will become the monitored object window 230 of the object of the item number "123", and thus the interface The user is provided to control the photographic picture of the selected particular item at each point in time, so the playback control unit 212 and the timeline control element 213 are no longer presented in a translucent manner, but are presented in a usable state.

In the monitoring environment window 210, except for the position A, a red frame will be presented, and the remaining positions (position B to position H) will present a translucent dark gray frame. At the same time, the monitoring environment window 210 presents a historical behavioral trajectory of the selected particular object. The historical behavior of a particular object can be obtained by analyzing and filtering. In more detail, first, the object information belonging to the object number "123" is obtained in the video and analysis data library 130. Then, the historical behavior track of the specific object is connected through the corresponding time and the spatial information that the object exists in the monitoring environment. In addition, in order to avoid multiple data capture and analysis, the user interaction platform 150 can temporarily store the object information of all the objects in the photographic picture currently being viewed.

4B is a detailed schematic diagram of a window of a monitored object according to an embodiment of the present invention. 4B is a detailed view of the monitored object window 230 of FIG. 4A. The display area 231 in the center of the monitoring object window 230 is used to present the current photographic image, and the current display area 231 also marks the camera number corresponding to the current photographic image. , shooting time and object information of each object. For example, the current photographic picture of FIG. 4B is taken by a camera No. 1, so that the display area 231 has the mark of the No. 1 camera.

In addition, the specific items selected by the user can be marked with a bold color (such as red color), while other objects will be marked with a non-obtrusive dotted line. The previous related object list 232 herein can be used to present a photographic picture of possible objects taken at a previous time and different cameras, and the photographic pictures of these possible objects will be in the previous related object list 232 according to the order of the object related probability. arrangement. The subsequent related object list 233 here can be used to present a photographic picture of possible objects taken after a few seconds of the current time and taken by different cameras, and the photographic pictures of these possible objects will be based on the object-related probability in the previously related object list 233. Arrange in high and low order.

For example, in the previous related object list 232 of FIG. 4B, the objects related to the object-related probability of the object corresponding to the object number "123" of the current photographic picture are sequentially appearing in the photographic image of the No. 3 camera. The object of the article number "123", the object of the article number "147" appearing in the photographic image of the camera No. 4, and the object of the article number "169" appearing in the photographic image of the camera No. 6. In this embodiment, the multi-video content analyzing unit 140 considers that the object of the object number "123" appearing in the photographic image of the camera No. 3 and the object of the object number "123" in the current photographic image should be the same object, The object number "123" appearing in the photographic image of camera No. 3, and the object number "147" appearing in the photographic image of camera No. 4 and the object number appearing in the photographic image of camera No. 6" The 169" object is indicated by a non-obscured dotted frame.

In addition, the photographic image presented in the previous related object list 232 may be a photographic image that may be completely present in the camera, or may be a partial photographic image of a possible object in the monitoring area of the camera, or may be The object appears in the photographic image of the superimposed behavior track in the monitoring area. In summary, the manner in which the previously associated object list 232 is used to present a photographic image of a possible object is not intended to limit the invention.

In addition, if there is a same object in the photographic image of the subsequent related object list 233 as the selected specific object, the photographic picture is arranged to the highest position. For example, because the monitoring area of camera No. 4 overlaps with the monitoring area of camera No. 1, the selected specific object (such as the object number "123") will appear in both the camera of the No. 1 camera and the No. 4 camera. In the picture. Since the photographed picture taken by the No. 4 camera contains the selected specific object (the object of the object number "123"), the photographed picture taken by the No. 4 camera is placed in the first priority position in the subsequent related object list 233. . At the same time, the selected objects in the photographic picture taken by the No. 4 camera will also be marked with a prominent object frame.

The previous object concatenation result 234 is used to present the photographic images of different cameras previously appearing on the selected object, and the photographic images are arranged in chronological order. The photographic image presented in the previous object serialization result 234 may be a photographic image in which the object may be completely presented in the camera, or may be a partial photographic image of the object in the monitoring area of the camera, or a possible object may appear in the monitoring. The photographic picture after the behavior track in the area is superimposed. In summary, the manner in which previous object serialization results 234 are presented is not intended to limit the invention.

Since FIG. 4B is the monitoring object window 230 in the case of real-time monitoring, the subsequent object serialization result 235 cannot know the future behavior trajectory of the selected specific object in the case of real-time monitoring, so the subsequent object serial connection result 235 It can still be rendered in dark color markings, or even in the monitoring object window 230.

If the user wants to see the photographic image of the previously selected specific object, the time axis control component 213 can be selected, or the playback control unit 212 can be used to view the selected object in the monitored object window 230 at a specified time. Photography screen. In other words, the user interaction platform 150 also has the function of performing a post-mortem review of a particular item.

When the user wants to view the photographic image of the specific time period through the user interaction platform 150, the interface presented on the user interaction platform 150 at this time includes the monitoring environment window 210, the camera list window 220, the playback control unit 212, and the time axis control. Element 213 and multi-camera photographic picture window 240. The multi-camera photographic picture window 240 presents a number of photographic pictures for a particular time period specified by the user, and these photographic pictures can be taken from the video and analytics data library 130. Each photographic picture may be presented by an independent sub-window or may be rendered by one of the divided pictures. As described above, the size and position of the independent sub-window can be set by the user, and the distribution pattern of each picture in the divided picture can also be defined by the user. The user can use the playback control unit 212 and the timeline control component 213 to simultaneously play or manipulate all of the photographic frames to thereby view the desired photographic images in the surveillance environment.

5A and 5B, FIG. 5A is a schematic diagram of an interface on a user interaction platform when a user selects a camera for post-mortem review according to an embodiment of the present invention, and FIG. 5B is a detailed schematic diagram of a specific camera monitoring window according to an embodiment of the present invention. The detailed schematic diagram of FIG. 5B corresponds to a specific camera monitoring window when the user selects the camera for post-mortem review.

When the user clicks on a particular camera from the multi-camera photographic screen 240, the camera distribution location in the monitoring environment window 210, or the camera list window 220, the particular camera monitoring window 250 is generated immediately. At the same time, in the monitoring environment window 210 and the camera list window 220, the selected cameras are marked in bold colors (such as red), while other unselected cameras are marked in non-obtrusive colors (such as dark gray). In addition, the multi-camera photographing screen window 240 will be reduced to the lower edge of the screen of the interface; or, the multi-camera photographing screen window 240 will be reduced to be located at the edge of the screen of the interface, and the photographing screens of other multi-cameras will be presented in a reduced screen.

The photographic screen currently being played by the selected camera will be presented on the display screen 231 of the particular camera monitoring window 250. In addition, the previous related object list 232 presents the played photographic picture of the selected camera adjacent to the camera a few seconds ago, and the subsequent related object list 233 presents the photographic picture played by the adjacent camera of the selected camera after a few seconds. In addition, since the user has not selected the item to be tracked, the previous object serialization result 234 and the subsequent object serial connection result 235 do not need to present any content, and may be presented in a dark color mark or not in a specific camera monitoring. In window 250.

For example, when the user clicks on the No. 1 camera, the specific camera monitoring window 250 corresponding to the No. 1 camera is immediately generated. At the same time, the camera No. 1 in the camera list window 220 will be presented in red, and the remaining cameras will be presented in dark gray. The position A in the environment diagram will present a red frame, and the remaining positions (position B to position H) will present a translucent dark gray frame. In addition, the multi-camera photographic screen window 240 will be reduced to the lower edge of the screen of the interface.

FIG. 6A is a schematic diagram of an interface on a user interaction platform when a user selects a specific object for post-mortem review according to an embodiment of the present invention. When the user clicks on a particular item, the particular camera monitoring window 250 will become the monitored object window 230. For example, after the user clicks on the object of the item number "123", the interface on the user interaction platform 150 will change, and the specific camera monitoring window 250 will become the monitored object window 230 of the object of the item number "123".

In the monitoring environment window 210, except for the position A, a red frame will be presented, and the remaining positions (position B to position H) will present a translucent dark gray frame. At the same time, the monitoring environment window 210 presents the historical behavior track of the selected specific object, wherein the circle marked by the monitoring environment window 210 indicates the position of the object in the monitoring environment, so the dot will change according to the position of the object during the playing time. And can be rendered in a blinking manner to highlight the location of the selected object in the monitored environment. Accordingly, the entire behavioral trajectory of the selected object will be obtained based on the results of the resulting tandem object.

FIG. 6B is a detailed schematic diagram of a window of a monitored object according to an embodiment of the present invention. 6B is a detailed view of the monitoring object window 230 in FIG. 6A. The display area 231 in the center of the monitoring object window 230 is used to present the current photographic picture, and the current display area 231 also marks the camera number corresponding to the current photographic picture. , shooting time and object information of each object. For example, the current photographic picture of FIG. 6B is taken by a camera No. 1, so that the display area 231 has the mark of the No. 1 camera.

The previous related object list 232 can be used to present photographic images of possible objects that were taken at a previous time and different cameras, and the photographic images of these possible objects will be ranked in the previous related object list 232 in accordance with the order of the object relevance ratings.

In addition, the photographic image presented in the previous related object list 232 may be a photographic image in which the object may be completely presented in the camera, or may be a partial photographic image of the object in the monitoring area of the camera, or a possible object may appear in the photographic image of the camera. The photographic picture superimposed on the behavior track in the monitoring area. In summary, the manner in which the previously associated object list 232 is used to present a photographic image of a possible object is not intended to limit the invention.

The subsequent related object list 233 here can be used to present a photographic picture of possible objects taken after a few seconds of the current playing time and taken by different cameras, and the photographic pictures of these possible objects will be based on the object relevance in the previously related object list 233. The order of the scores is arranged in order.

The previous object concatenation result 234 is used to present the photographic pictures in the monitored areas of the different cameras previously appearing by the selected object, and the photographic pictures are arranged in chronological order. The photographic image presented in the previous object serialization result 234 may be a photographic image in which the object may be completely presented in the camera, or may be a partial photographic image of the object in the monitoring area of the camera, or a possible object may appear in the monitoring. The photographic picture after the behavior track in the area is superimposed.

The subsequent object concatenation result 235 is used to present the photographic pictures in the monitoring area of different cameras that appear after the current time of the selected object, and the photographic pictures are arranged in chronological order. The photographic image presented in the subsequent object serialization result 235 may be a photographic image in which the object may be completely presented in the camera, or may be a partial photographic image of the object in the monitoring area of the camera, or a possible object may appear in the monitoring. The photographic picture after the behavior track in the area is superimposed.

FIG. 7 is a detailed schematic diagram of a monitored object window when a plurality of camera monitoring systems are connected in series with an error according to an embodiment of the present invention. In Fig. 7, it can be seen that the multi-camera monitoring system 100 has an error occurring when the object number "123" is concatenated. During the operation of the instant monitoring or the post-mortem review, the multi-video content analyzing unit 140 may cause the object to be concatenated incorrectly, so that different objects are recognized as the same object, and the user interaction platform 150 is presented. Smooth but erroneous object trajectory information and historical images.

When the user views the trajectory information and the historical image of the selected object, it is possible to find that different objects are marked as the same object number. In the embodiment of FIG. 7, the object with the item number "123" is the specific item that the user selects to track. In the monitored object window 230, the display area 231 displays the screen of the camera No. 1 with the shooting time of 12:06:30, and since the object whose object number is "123" is selected, the multi-video content analyzing unit 140 will object to the object. The object numbered "123" is connected in series.

In this embodiment, the object whose object number is "123" is substantially a personnel, but the multi-video content analysis unit 140 mistakes the B personnel for the object whose object number is "123", and the wrong serial result is generated. . Accordingly, the photographic image presented by the subsequent object serialization result 235 is not the correct behavior trajectory of the personnel.

At this time, the user only has to select the photographic picture of the correct object determined by the user from the subsequent related object list 233. In this embodiment, the user will select the photographic image of the object whose number is "126" recorded by the 12th camera at the shooting time of 12:06:40. Next, the display area 231 will display the photographic screen selected by the user in the subsequent object list. After the user selects the object with the item number "126" (essentially a person), the interface will display a confirmation message asking whether to correct. After the user confirms, the interface will transmit the correction data to the multi-video content analysis unit 140, and the multi-video content analysis unit 140 corrects the object information according to the correction data, that is, the object number and the object number whose object number is "123" is The object of "126" is re-aligned after the time of 12:06:40, and then the result of the serial connection is corrected, so that the personnel are marked as the object with the object number "123", and the personnel are all Mark the object with the item number "126". In addition, the corrected concatenation result will be stored in the video and analysis data repository 130 in addition to notifying the user interaction platform 150.

When the splicing result is corrected, the photographic image selected by the user is not the photographic image with the highest probability value, so the user interaction platform 150 will notify the analyzing unit 140 that the specific object currently being tracked should appear on the No. 12 camera. The shooting time is 12:06:40 and the picture number of the object with the item number "126" is taken. In addition, the multi-video content analyzing unit 140 compares the object information in the photographic image selected by the user according to the object feature of the currently tracked object and related object information, and gives the suggested splicing object displayed by using the red virtual frame. . If the user determines that the suggested splicing object is the correct object, the user only needs to click the red imaginary box, and the user interaction platform 150 sends the correction data to the analyzing unit 140. If the user thinks that the proposed splicing object is an erroneous object, the user can click on other objects indicated by the imaginary box. After the user clicks on the object indicated by the other virtual frame, the interface will again ask the user whether to confirm the correction, and after the user confirms the calibration, the user interaction platform 150 will send the correction data to the multi-video content analysis unit. 140.

After the interface used in the object serial connection correction method of the photographing screen provided by the embodiment of the present invention is described in detail, each step of the object serial connection correcting method of the photographing screen will be described next using a flowchart. Please refer to FIG. 8. FIG. 8 is a flowchart of a method for correcting serial connection of objects in a photographing screen according to an embodiment of the present invention. First, in step S800, a photographing screen of each camera in the multi-camera monitoring system is acquired. Next, in step S801, each photographic image is analyzed to obtain information about each object in each photographic image, wherein the object information includes the object number, the object feature, and the object type.

Then, in step S802, a user interaction platform is provided to the user, so that the user selects a specific object to be tracked through the user interaction platform. In step S803, the multi-camera monitoring system calculates the correlation between the object of the photographing screen of each camera before the photographing time of the current photographing screen and the specific object. In step S804, the multi-camera monitoring system calculates the correlation between the object of the photographing screen of each camera and the specific object after the photographing time of the current photographing screen.

In step S805, the multi-camera monitoring system automatically concatenates the specific objects appearing on each photographic image to obtain trajectory information and historical images of the specific object, wherein the manner of automatically splicing the specific object is to associate the specific object with the related object. The highest rated object is concatenated.

In step S806, the photographic images of the objects are sequenced according to the previous related object list of the user interaction platform according to the correlation between the photographic images of the cameras before the shooting time of the current photographic image and the specific objects. In step S807, the photographic images of the objects are sequenced according to the subsequent related object list of the user interaction platform according to the correlation between the photographic images of the cameras and the specific objects after the shooting time of the current photographic screen.

In step S808, the trajectory information of the specific object automatically connected in series and the historical image before the shooting time of the current photographic image, and the photographic images of the specific object captured by the non-current camera are sequentially arranged in use. The previous object of the interactive platform is concatenated in the results. In step S809, the trajectory information of the specific object automatically connected in series and the shooting time of the current photographic image in the historical image, and the plurality of photographic images of the specific object captured by the current camera are sequentially arranged. The subsequent object of the user interaction platform is concatenated in the result.

If the use of the found automatic cascading result is wrong, the user will click on the correct spliced object in the photographic image that is not the most relevant in the subsequent object list to correct. According to this, in step S810, it is judged whether or not the one photographing screen whose relevance in the subsequent object list is not the largest is clicked. If the photographic image whose relevance is not the largest in the subsequent object list is not selected, it indicates that the automatic cascading result of the object is the correct cascading result, and the object serial connection correction method of the photographic image is ended.

If the photographic image whose relevance is not the largest in the subsequent object list is clicked, in step S811, the selected photographic image is displayed as the current photographic image, and the user selects the object of the current photographic image. After that, ask the user whether to correct the result of the automatic serial connection of the object. If the user does not correct the result of the automatic serial connection of the object, the object serial connection correction method of the photographing screen is ended. If the user confirms that the result of the automatic serial connection of the object is corrected, then in step S812, the user interaction platform generates correction data according to the selected object of the photographic picture to the multi-camera monitoring system, so that the multi-camera monitoring system generates suggestions. The object is concatenated with the correction result.

Thereafter, in step S813, the user interaction platform asks the user whether the suggested object serialization correction result is the correct concatenation result of the specific object. If the user thinks that the suggested object serialization correction result is the correct concatenation result of the specific object, then in step S814, the suggested object serialization correction result is taken as the correct concatenation result of the specific object, and then the object of the photographic image is ended. Tandem correction method. Conversely, if the user does not consider the suggested object serialization correction result as the correct object serialization result, the process returns to step S810.

In summary, the multi-camera monitoring system provided by the embodiment of the present invention has a method for correcting the object serial connection of the photographic image, and the multi-camera monitoring system has a user interaction platform for the user to operate, and the user performs the photographic image. The object serial connection correction method is to correct the error that may occur in the automatic serial connection object of the conventional multi-camera monitoring system.

Although the preferred embodiments of the present invention have been disclosed as above, the present invention is not limited to the above-described embodiments, and any one of ordinary skill in the art can make some modifications without departing from the scope of the present invention. The scope of protection of the present invention should be determined by the scope of the appended claims.

100. . . Multi-camera monitoring system

110. . . Video capture analysis unit

120. . . Video analysis data collection unit

130. . . Video and analytics data library

140. . . Multi-video content analysis unit

150. . . User interaction platform

210. . . Monitoring environment window

211. . . Environmental schematic

212. . . Playback control unit

213. . . Time axis control element

210. . . Camera list window

230. . . Monitor object window

231. . . Display area

232. . . Previous object list

233. . . Follow-up object list

234. . . Previous object concatenation result

235. . . Subsequent object concatenation results

240. . . Multi-camera photographic screen

241. . . Unique window

242. . . Split screen

250. . . Specific camera monitoring window

S800～S814. . . Step flow

FIG. 1 is a block diagram of a multi-camera security monitoring system according to an embodiment of the present invention.

2 is a schematic diagram of an interface of a method for correcting object serial connection of a photographic image on a user interaction platform according to an embodiment of the present invention.

FIG. 3A is a schematic diagram of an interface on a user interaction platform when a user selects a camera for real-time monitoring according to an embodiment of the present invention.

FIG. 3B is a detailed schematic diagram of a specific camera monitoring window according to an embodiment of the present invention.

FIG. 4A is a schematic diagram of an interface on a user interaction platform when a user selects a specific object for real-time monitoring according to an embodiment of the present invention.

4B is a detailed schematic diagram of a window of a monitored object according to an embodiment of the present invention.

FIG. 5A is a schematic diagram of an interface on a user interaction platform when a user selects a camera for post-mortem review according to an embodiment of the present invention.

FIG. 5B is a detailed schematic diagram of a specific camera monitoring window according to an embodiment of the present invention.

FIG. 6A is a schematic diagram of an interface on a user interaction platform when a user selects a specific object for post-mortem review according to an embodiment of the present invention.

FIG. 6B is a detailed schematic diagram of a window of a monitored object according to an embodiment of the present invention.

FIG. 7 is a detailed schematic diagram of a monitored object window when a plurality of camera monitoring systems are connected in series with an error according to an embodiment of the present invention.

FIG. 8 is a flowchart of a method for correcting object serial connection of a photographing screen according to an embodiment of the present invention.

230. . . Monitor object window

231. . . Display area

232. . . Previous object list

233. . . Follow-up object list

234. . . Previous object concatenation result

235. . . Subsequent object concatenation results

Claims (8)

A method for serially connecting a photographic image to form an object trajectory includes: selecting an object to be tracked through a user interaction platform; indicating a first plurality of captured by a multi-camera monitoring system during a recording time a series of images, wherein each of the serial images has a degree of correlation with the object being tracked; and a second plurality of strings are coupled from the first plurality of serial images according to the degree of correlation with the object being tracked And connecting the image to generate a first object serial connection result; and if an incorrect serial image is found in the first object serial connection result, selecting a serial image through the user interaction platform to replace the wrong serial connection And generating, by the image, a second serial connection result, and updating, according to the selected serial image, the serial image after the erroneous serial image in the first object serial connection result, wherein the selected serial connection The extent to which the image is related to the object being tracked is less than the degree to which the erroneous tandem image is related to the object being tracked. The method of claim 1, wherein the erroneous concatenated image and the selected concatenated image are captured by different cameras. The method of claim 1, wherein the first plurality of concatenated images comprises a previous concatenated image and a subsequent concatenated image, wherein the user interaction platform includes a monitoring window for presenting the current monitoring The object, the previously concatenated image, and the subsequent concatenated image. The method of claim 1, further comprising: searching for objects that may be met in each photographic image according to the tracking object to analyze the photographic images; and analyzing the multi-video content of the multi-camera monitoring system. The unit calculates the correlation between the object of the current photographic picture and the particular object. The method of claim 1, wherein the first serial connection result and the selected serial image are simultaneously viewed through the user platform to determine the erroneous serial image and the selected serial image. . The method of claim 3, wherein the user interaction platform further comprises: a monitoring environment window, including an environment diagram, wherein the environment diagram is used to present an overall monitoring environment of the multi-camera monitoring system, wherein The monitoring environment includes a geographic characteristic of the monitoring environment, a distribution location of each camera, and a behavior track of a specific object in the monitoring environment; a camera list window is used to present each camera number and the distribution of each camera in the monitoring environment. a relationship between the locations; and a multi-camera photographic window for presenting an instant photographic image captured by a plurality of cameras selected by the user, or for playing a plurality of selected cameras recorded in a database Historical video material. The method of claim 6, wherein the monitoring environment window further comprises: a playback control unit, configured to effectively control the playback of the video data when tracking, presenting and correcting the subsequent historical track of the specific object. And a time axis control component for controlling the video data to be played back and forth at a specific time point. The method of claim 6, wherein the environment diagram is one selected from a geographic environment map, an architectural architecture map, and a monitoring facility distribution map, or a stack of the selected ones or all of the maps. And at least one of the plurality of figures is superimposed through the three-dimensional computer image.