A video linkage system based on large-scale video surveillance

Abstract

The present invention discloses a video linkage system based on large-scene video monitoring. The system analyzes the orientation of all cameras of a controlled scene through a viewing direction analysis unit to obtain their orientation positions, and after determining the coverage area according to the orientation position, forms a standard architectural model of the controlled scene; then, the controlled video group is initially spliced ​​by a cutting and splicing unit in combination with the architectural model, and all controlled videos are initially fused to obtain an initially fused video, and the intersection area is processed in a preliminary manner; then, the cutting and splicing unit is used to transmit the architectural model and the initially fused video to a picture direction analysis unit via a processor, and the picture analysis unit is used to perform a two-way analysis on the real-time initially fused video in combination with the picture direction splicing unit, and the display screen of the intersection area is displayed in the manner provided by the present application, and abnormal conditions are monitored in real time, thereby providing a reasonable and efficient joint monitoring system; the present invention is simple, effective, and easy to use.

Description

Technical Field

[0001] The present invention relates to the technical field of video linkage, and in particular to a video linkage system based on large-scene video monitoring. Background Art

[0002] The patent with publication number CN112770088A discloses an AI video linkage perception monitoring system, including: a visual module, an industrial data acquisition system, an AI video linkage perception monitoring platform and a display terminal; the visual module is used to identify and collect videos and upload them to the AI ​​video linkage perception monitoring platform; the industrial data acquisition module is used to collect real-time structured data of mechanical equipment and MES and upload them to the AI ​​video linkage perception monitoring platform; the AI ​​video linkage perception monitoring platform is used to receive the alarm data sent by the industrial data acquisition module, and obtain the source area of ​​the alarm data, and then call the visual module of the source area for identification and capture; it is also used to train the algorithm model, and call the algorithm to load it into the visual module to perform intelligent analysis and recognition of the video. It can realize the digital display of abnormal situations through artificial intelligence analysis and linkage equipment data.

[0003] However, currently, when monitoring large scenes, due to some scene restrictions, a single camera cannot cover the entire area, and a linked monitoring network consisting of many cameras is required to monitor the entire area. However, there is a problem with this monitoring, that is, how to display the cross-coverage area and which direction to focus on, especially when performing video splicing and fusion monitoring. Based on this, the present application provides a solution. Summary of the invention

[0004] The purpose of the present invention is to provide a video linkage system based on large-scale video surveillance;

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] A video linkage system based on large-scale video surveillance, characterized in that it includes

[0007] The cropping and splicing unit is used to transmit the building model and the initial fusion video to the image analysis unit via the processor. The image analysis unit is used to perform a two-way analysis on the real-time initial fusion video in combination with the image splicing unit. The specific method of the two-way analysis is as follows:

[0008] SS1: Get all the intersection areas in the initial fusion video;

[0009] SS2: Choose any intersection area;

[0010] SS3: Obtain all controlled videos contained in the intersection area and mark them as sub-item videos Fi, where i=1...n, indicating that there are n sub-item videos;

[0011] SS4: Then, face capture is performed on all sub-item videos Fi to obtain the number of faces in each sub-item video, and the face is marked as a direction value Ci, i=1...n; Ci and Fi are in a one-to-one correspondence;

[0012] SS5: Then automatically calculate the sum of all orientation values ​​Ci and mark it as the total number of orientation faces;

[0013] SS6: Calculate the time-sharing ratio Zi, i = 1...n; the calculation formula is:

[0014] Zi = Ci / total number of faces;

[0015] SS7: Then, T2 is automatically multiplied by Zi to obtain the display time of each sub-item video; according to the specific value of the display time, the display time length of the corresponding sub-item video in the intersection area is adjusted by using the picture-wise splicing unit;

[0016] SS8: Select the next intersection area and repeat steps SS3-SS8 to complete the processing of all intersection areas to obtain a new video, which is marked as a nuclear fusion video;

[0017] The picture-to-picture splicing unit is used to transmit the nuclear fusion video to the processor, and the processor is used to transmit the nuclear fusion video to the abnormality determination unit, and the abnormality determination unit is used to perform abnormality analysis on the nuclear fusion video, specifically:

[0018] SS01: Get the nuclear fusion video. When the vertical height of a person suddenly decreases in the video, the vertical height is calculated by the conversion formula between pixel points and distance. It can also be directly replaced by pixel points. When the vertical height decreases by more than X1 value within T3 time, an early warning signal is generated. T3 and X1 are both preset values.

[0019] SS02: Continuously monitor the personnel who generate the warning signal. If the vertical height of the personnel has not recovered to the height before T4+T3 at T4, an alarm signal will be generated and the corresponding personnel will be marked as an alarm personnel; T3 and T4 are both preset values, and T4≥10T3;

[0020] SS03: If the alarmed person is in the intersection area, all sub-item videos that can cover the intersection area are obtained and combined to form a hazard control video group.

[0021] Furthermore, it also includes a video acquisition unit, which is used to acquire all controlled videos of the controlled scene. All controlled videos constitute a controlled video group, and the controlled video has a location identifier for identifying which specific camera has acquired it; the controlled scene is the monitored large scene video; the video acquisition unit is used to transmit the controlled video group to the cropping and splicing unit.

[0022] Furthermore, it also includes a viewing direction analysis unit;

[0023] A viewing direction analysis unit is used to analyze the directions of all cameras in the controlled scene and obtain a building model containing the corresponding coverage area of ​​the cameras;

[0024] The viewing direction analysis unit is used to transmit the building model to the cutting and splicing unit, and the cutting and splicing unit receives the building model transmitted by the viewing direction analysis unit;

[0025] The cropping and splicing unit receives the controlled video group transmitted by the video acquisition unit; the cropping and splicing unit receives the building model transmitted by the viewing direction analysis unit, and is used to perform initial splicing on the controlled video group in combination with the building model. The specific method of initial splicing is:

[0026] S01: Get the building model;

[0027] S02: Afterwards, the coverage areas of all cameras in the building model are obtained, and the overlapping parts of the coverage areas of all cameras are obtained and marked as intersection areas;

[0028] S03: Obtain all controlled videos, and preliminarily splice all controlled videos according to the method in the building model;

[0029] S04: Afterwards, all the intersection areas are obtained, and one of the intersection areas is selected to obtain the number of overlapped controlled videos in the intersection area, and the overlapped number is marked as an overlap value, and T2 is divided by the overlap value to obtain the carousel time; T2 is the time value preset by the administrator;

[0030] The above processing is performed on all the intersection areas simultaneously to obtain the carousel of all the intersection areas;

[0031] S05: All controlled videos are merged, and then all controlled videos in the crossover area are played in turn according to the carousel; an initial merged video is obtained, and the initial merged video includes a plurality of spliced ​​controlled videos.

[0032] Furthermore, the specific method of orientation analysis is as follows:

[0033] Step 1: Obtain the building model of the controlled scene, which includes the environment of the controlled scene, the project layout of each location, and the specific layout of related dimensions;

[0034] Step 2: Then mark the position of each camera in the controlled scene and calibrate the direction at the same time. The specific method of direction calibration is as follows:

[0035] First, the main direction analysis is performed to obtain the position movement of all cameras. A single movement refers to the change in the direction of the video obtained by the camera, and after stopping, it does not change again within T1 time, and it is marked as a movement point; T1 is a preset value;

[0036] The duration of each corresponding moving point is obtained, and the duration is the time from the last moving to the corresponding moving point or the start of setting the position to the next moving point, which is marked as the duration of the corresponding moving point;

[0037] Then, the tail distance duration corresponding to each moving point is obtained. The tail distance duration is the time from the current position of the moving point to the current position after it is determined. This time is the tail distance duration;

[0038] Then the exact value of the moving point is automatically calculated. The specific calculation formula is:

[0039] Confirmation value = 0.75*duration + 0.25 / tail distance duration;

[0040] Obtain the confirmation values ​​of all moving points, mark the moving point with the largest confirmation value as the main direction, and mark the area corresponding to the main direction as the coverage area of ​​the current camera, and within the building model where the controlled scene is located;

[0041] Step 3: Update the building model.

[0042] Furthermore, it also includes an abnormality determination unit;

[0043] The abnormality determination unit is used to transmit the danger control video group and the alarm personnel to the prompt unit, and the prompt unit is used to remind the management personnel to handle it.

[0044] Further, it also includes a management unit;

[0045] The management unit is in communication with the processor and is used to input all preset values;

[0046] The processor is also used to transmit the nuclear fusion video to the storage unit for storage backup every day.

[0047] Beneficial effects of the present invention:

[0048] The present invention uses a viewing direction analysis unit to analyze the orientation of all cameras in the controlled scene, obtains their orientation positions, and determines the coverage area according to the orientation positions to form a standard architectural model of the controlled scene; then, a cropping and splicing unit is used to perform initial splicing of the controlled video group in combination with the architectural model, and all controlled videos are initially fused to obtain an initial fused video, and the cross-region is processed in a preliminary manner;

[0049] Afterwards, the cropping and splicing unit is used to transmit the building model and the initial fusion video to the picture analysis unit via the processor. The picture analysis unit is used to perform a two-way analysis on the real-time initial fusion video in combination with the picture splicing unit, and display the display screen of the intersection area in the manner provided by the present application, and monitor abnormal situations in real time, thereby providing a reasonable and efficient joint monitoring system. The present invention is simple, effective, and easy to use. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The present invention will be further described below in conjunction with the accompanying drawings.

[0051] Figure 1 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] See also Figure 1 As shown, the present invention is a video linkage system based on large-scale video surveillance, the system comprising a video acquisition unit, a viewing direction analysis unit, a cropping and splicing unit, a processor, a management unit, a storage unit, a viewing direction analysis unit, a viewing direction splicing unit, an abnormality determination unit and a prompt unit;

[0054] The video acquisition unit is used to acquire all controlled videos of the controlled scene, and all controlled videos constitute a controlled video group, and the controlled video has a position mark for identifying which camera specifically acquires it; the controlled scene is a large scene video being monitored, and the scene is large, and a single video cannot cover it all, and there is a covered area, and a single camera cannot meet the real-time monitoring of all scenes, so multiple cameras are required for coordinated monitoring, but the monitoring camera angles set for each scene are different, and there will be overlap when covering, but the conversion may be troublesome when involving other cameras;

[0055] The video acquisition unit is used for transmitting the controlled video group to the cutting and splicing unit;

[0056] The viewing direction analysis unit is used to analyze the direction of all cameras in the controlled scene and obtain their direction positions. The specific method of direction analysis is as follows:

[0057] Step 1: Obtain the building model of the controlled scene, which includes the environment of the controlled scene, the project layout of each location, and the specific layout of related dimensions;

[0058] Step 2: Then mark the position of each camera in the controlled scene and calibrate the direction at the same time. The specific method of direction calibration is as follows:

[0059] First, the main direction analysis is performed to obtain the position movement of all cameras. A single movement refers to the change in the direction of the video obtained by the camera, and after stopping, it does not change again within T1 time, and it is marked as a movement point; T1 is a preset value;

[0060] The duration of each corresponding moving point is obtained, and the duration is the time from the last moving to the corresponding moving point or the start of setting the position to the next moving point, which is marked as the duration of the corresponding moving point;

[0061] Then, the tail distance duration corresponding to each moving point is obtained. The tail distance duration is the time from the current position of the moving point to the current position after it is determined. This time is the tail distance duration;

[0062] Then the exact value of the moving point is automatically calculated. The specific calculation formula is:

[0063] Confirmation value = 0.75*duration + 0.25 / tail distance duration;

[0064] Obtain the confirmation values ​​of all moving points, mark the moving point with the largest confirmation value as the main direction, and mark the area corresponding to the main direction as the coverage area of ​​the current camera, and within the building model where the controlled scene is located;

[0065] Step 3: Update the building model;

[0066] The viewing direction analysis unit is used to transmit the building model to the cutting and splicing unit, and the cutting and splicing unit receives the building model transmitted by the viewing direction analysis unit;

[0067] The cropping and splicing unit receives the controlled video group transmitted by the video acquisition unit; the cropping and splicing unit receives the building model transmitted by the viewing direction analysis unit, and is used to perform initial splicing on the controlled video group in combination with the building model. The specific method of initial splicing is:

[0068] S01: Get the building model;

[0069] S02: Afterwards, the coverage areas of all cameras in the building model are obtained, and the overlapping parts of the coverage areas of all cameras are obtained and marked as intersection areas;

[0070] S03: Obtain all controlled videos, and preliminarily splice all controlled videos according to the method in the building model;

[0071] S04: Afterwards, all the intersection areas are obtained, and one of the intersection areas is selected to obtain the number of overlapped controlled videos in the intersection area, and the overlapped number is marked as an overlap value, and T2 is divided by the overlap value to obtain the carousel time; T2 is the time value preset by the administrator;

[0072] The above processing is performed on all the intersection areas simultaneously to obtain the carousel of all the intersection areas;

[0073] S05: All controlled videos are merged, and then all controlled videos in the crossover area are played in turn according to the carousel; an initial merged video is obtained, and the initial merged video includes a plurality of spliced ​​controlled videos;

[0074] The cropping and splicing unit is used to transmit the building model and the initial fusion video to the image analysis unit via the processor. The image analysis unit is used to perform a two-way analysis on the real-time initial fusion video in combination with the image splicing unit. The specific method of the two-way analysis is as follows:

[0075] SS1: Get all the intersection areas in the initial fusion video;

[0076] SS2: Choose any intersection area;

[0077] SS3: Obtain all controlled videos contained in the intersection area and mark them as sub-item videos Fi, where i=1...n, indicating that there are n sub-item videos;

[0078] SS4: Then, face capture is performed on all sub-item videos Fi to obtain the number of faces in each sub-item video, and the face is marked as a direction value Ci, i=1...n; Ci and Fi are in a one-to-one correspondence;

[0079] SS5: Then automatically calculate the sum of all orientation values ​​Ci and mark it as the total number of orientation faces;

[0080] SS6: Calculate the time-sharing ratio Zi, i = 1...n; the calculation formula is:

[0081] Zi = Ci / total number of faces;

[0082] SS7: Then, T2 is automatically multiplied by Zi to obtain the display time of each sub-item video; according to the specific value of the display time, the display time length of the corresponding sub-item video in the intersection area is adjusted by using the picture-wise splicing unit;

[0083] SS8: Select the next intersection area and repeat steps SS3-SS8 to complete the processing of all intersection areas to obtain a new video, which is marked as a nuclear fusion video;

[0084] The picture-to-picture splicing unit is used to transmit the nuclear fusion video to the processor, and the processor is used to transmit the nuclear fusion video to the abnormality determination unit, and the abnormality determination unit is used to perform abnormality analysis on the nuclear fusion video, specifically:

[0085] SS01: Get the nuclear fusion video. When the vertical height of a person suddenly decreases in the video, the vertical height is calculated by the conversion formula between pixel points and distance. It can also be directly replaced by pixel points. When the vertical height decreases by more than X1 value within T3 time, an early warning signal is generated. T3 and X1 are both preset values.

[0086] SS02: Continuously monitor the personnel who generate the warning signal. If the vertical height of the personnel has not recovered to the height before T4+T3 at T4, an alarm signal will be generated and the corresponding personnel will be marked as an alarm personnel; T3 and T4 are both preset values, and T4≥10T3;

[0087] SS03: If the alarmed person is in the intersection area, all sub-item videos covering the intersection area are obtained and combined to form a hazard control video group;

[0088] The abnormality determination unit is used to transmit the danger control video group and the alarm personnel to the prompt unit, and the prompt unit is used to remind the management personnel to handle it.

[0089] The management unit is in communication with the processor and is used to input all preset values;

[0090] The processor is also used to transmit the nuclear fusion video to the storage unit for storage backup every day.

[0091] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A video linkage system based on large-scale video surveillance, characterized in that: include: It includes a video acquisition unit, which is used to acquire all controlled videos of the controlled scene. All controlled videos constitute a controlled video group, and the controlled video has a position mark, which is used to identify which camera specifically acquires it; the controlled scene is the monitored large scene video; the video acquisition unit is used to transmit the controlled video group to the cutting and splicing unit; It also includes a viewing direction analysis unit; A viewing direction analysis unit is used to analyze the directions of all cameras in the controlled scene and obtain a building model containing the corresponding coverage area of ​​the cameras; The viewing direction analysis unit is used to transmit the building model to the cutting and splicing unit, and the cutting and splicing unit receives the building model transmitted by the viewing direction analysis unit; The cropping and splicing unit receives the controlled video group transmitted by the video acquisition unit; The cropping and splicing unit receives the building model transmitted by the viewing direction analysis unit. The cropping and splicing unit is used to perform initial splicing of the controlled video group in combination with the building model. The specific method of initial splicing is: S01: Get the building model; S02: Afterwards, the coverage areas of all cameras in the building model are obtained, and the overlapping parts of the coverage areas of all cameras are obtained and marked as intersection areas; S03: Obtain all controlled videos, and preliminarily splice all controlled videos according to the method in the building model; S04: Afterwards, all the intersection areas are obtained, and one of the intersection areas is selected to obtain the number of overlapped controlled videos in the intersection area, and the overlapped number is marked as an overlap value, and T2 is divided by the overlap value to obtain the carousel time; T2 is the time value preset by the administrator; The above processing is performed on all the intersection areas simultaneously to obtain the carousel of all the intersection areas; S05: All controlled videos are merged, and then all controlled videos in the crossover area are played in turn according to the carousel; an initial merged video is obtained, and the initial merged video includes a plurality of spliced ​​controlled videos; The cropping and splicing unit is used to transmit the building model and the initial fusion video to the image analysis unit via the processor. The image analysis unit is used to perform a two-way analysis on the real-time initial fusion video in combination with the image splicing unit. The specific method of the two-way analysis is as follows: SS1: Get all the intersection areas in the initial fusion video; SS2: Choose any intersection area; SS3: Get all controlled videos contained in the intersection area and mark them as sub-item videos Fi, where i=1...n, indicating that there are n sub-item videos; SS4: Then, face capture is performed on all sub-item videos Fi to obtain the number of faces in each sub-item video, which is marked as the orientation value Ci, i=1...n; Ci and Fi are in one-to-one correspondence; SS5: Then automatically calculate the sum of all orientation values ​​Ci and mark it as the total number of orientation faces; SS6: Calculate the time-sharing ratio Zi, i=1...n; the calculation formula is: Zi=Ci / total number of faces; SS7: Then, T2 is automatically multiplied by Zi to obtain the display time of each sub-item video; according to the specific value of the display time, the display time length of the corresponding sub-item video in the intersection area is adjusted by using the picture-wise splicing unit; SS8: Select the next intersection area and repeat steps SS3-SS7 to complete the processing of all intersection areas to obtain a new video, which is marked as a nuclear fusion video; The picture-wise stitching unit is used to transmit the nuclear fusion video to the processor, and the processor is used to transmit the nuclear fusion video to the abnormality determination unit, and the abnormality determination unit is used to perform abnormality analysis on the nuclear fusion video to obtain a hazard control video group.

2. According to claim 1, a video linkage system based on large-scene video surveillance is characterized in that: The specific method of the abnormality determination unit performing abnormality analysis on the nuclear fusion video to obtain the risk control video group is: SS01: When the nuclear fusion video is obtained, when the vertical height of a person in the video suddenly decreases, the vertical height is calculated by the conversion formula between pixel points and distance. When the vertical height decreases by more than the X1 value within T3 time, an early warning signal is generated; T3 and X1 are both preset values; SS02: Continuously monitor the personnel who generate the warning signal. If the vertical height of the personnel has not recovered to the height before T4+T3 at T4, an alarm signal will be generated and the corresponding personnel will be marked as an alarm personnel; T3 and T4 are both preset values, and T4≥10T3; SS03: If the alarmed person is in the intersection area, all sub-item videos that can cover the intersection area are obtained and combined to form a hazard control video group.

3. A video linkage system based on large-scene video surveillance according to claim 2, characterized in that: The abnormality determination unit is used to transmit the danger control video group and the alarm personnel to the prompt unit, and the prompt unit is used to remind the management personnel to handle it.

4. The video linkage system based on large-scene video surveillance according to claim 1 is characterized in that: Also includes management unit; The management unit is in communication with the processor and is used to input all preset values; The processor is also used to transmit the nuclear fusion video to the storage unit for storage backup every day.

Images