A 2D visualization-based video inspection method

Abstract

The application discloses a kind of video inspection methods based on 2D visualization, comprising: calling inspection configuration library, parsing and displaying the configuration distribution diagram and inspection route in inspection configuration library;Select the video inspection of inspection route corresponding to configuration distribution diagram, and inspection route includes camera sequence list and camera control parameter;According to the real-time video of camera sequence list calling camera is played, and according to camera control parameter control camera video acquisition;When the real-time video of multiple cameras needs to be multicast, pop up several windows, and each window corresponds to play the real-time video of a camera.The application fully utilizes camera resources to play video inspection by the setting of camera sequence list and camera control parameter in inspection route, while intuitively displaying the position of camera, monitoring area and inspection route in configuration distribution diagram;And configuration distribution diagram and inspection route are stored in inspection configuration library corresponding to facilitate calling and modification.

Description

Technical Field

[0001] This invention relates to the field of inspection technology, and in particular to a video inspection method based on 2D visualization. Background Technology

[0002] The inspection task is the aggregate carrier of inspection rules and provides the function of modifying the execution scope and scheduling method of the rules. The existing inspection process includes manual inspection and camera switching inspection. However, when the inspection area is large, the inspection route is diverse, and the inspection frequency is high, manual inspection is costly in terms of manpower and has low work efficiency. When using the pre-set camera switching inspection, each camera in the existing camera monitoring interface plays independently. Even when multiple cameras play in parallel, the screen will be smaller due to splitting. The playback configuration mode is limited and cannot make full use of camera resources. At the same time, it cannot intuitively display the actual location information of each camera and the location of the monitoring area, and the inspection personnel cannot understand the specific inspection route.

[0003] A Chinese patent document, CN110830763A, published on February 21, 2020, entitled "A Method and Apparatus for Surveillance Video Inspection," includes: receiving an inspection request sent by a video network terminal; inspecting surveillance videos of monitoring points within a preset area according to the inspection request; and sending the inspection results to the video network terminal so that the video network terminal can control the monitoring points in the preset area where the surveillance videos have been successfully played to be highlighted on the display interface. The inspection results include at least information about the monitoring points where the surveillance videos have been successfully played, as determined by the video network server. In this technology, the video network terminal can adjust the display of monitoring points based on the inspection results of the video network server, allowing users to clearly understand which surveillance videos are viewable, thus improving the efficiency of viewing surveillance videos. However, this technology only improves the efficiency of switching between viewing surveillance videos; it does not address the problem of the single configuration mode of camera playback, cannot fully utilize camera resources, and cannot determine the actual location and monitoring area of ​​the cameras, nor can it display inspection route information. Summary of the Invention

[0004] This invention aims to overcome the problems of existing technologies, such as the single configuration mode of camera playback, insufficient resource utilization, and inability to intuitively display inspection route information and camera monitoring areas. It provides a 2D visualization-based video inspection method that fully utilizes camera resources for video inspection playback by setting the camera sequence list and camera control parameters in the inspection route. At the same time, the location of the cameras, the monitoring area, and the inspection route can be intuitively displayed in the configuration distribution map. The configuration distribution map and the inspection route are stored in the corresponding inspection configuration library for easy retrieval and modification.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A video inspection method based on 2D visualization includes:

[0007] Retrieve the inspection configuration library, parse and display the configuration distribution map and inspection route in the inspection configuration library;

[0008] Select the inspection route corresponding to the configuration distribution map for video inspection. The inspection route includes a camera sequence list and camera control parameters.

[0009] The system retrieves and plays the live video from the corresponding camera according to the camera sequence list, and controls the video capture of the camera playing the video according to the camera control parameters. When it is necessary to multicast the live video from multiple cameras, several windows pop up, each window playing the live video from one camera.

[0010] This invention primarily addresses situations involving large inspection areas, numerous inspection routes, and high inspection frequencies. It binds relevant information from actual cameras to corresponding cameras on a configuration distribution map. Users can draw inspection routes on the map, which automatically include all surrounding cameras, forming an initial camera sequence list and corresponding camera control parameters. Users can modify the inspection routes, including camera rotation speed and viewing angle dwell time. After configuration, the configuration distribution map and inspection routes are stored in the inspection configuration library for later retrieval. Users can choose to play a customized inspection route. Through a video playback component, the actual cameras are retrieved according to the configuration distribution map and inspection route, and real-time video is displayed on the playback interface, enabling online inspection tasks.

[0011] Preferably, the process of building the inspection configuration library includes:

[0012] Obtain the configuration distribution map of the area to be inspected, including the location, orientation, and field of view coverage of the cameras;

[0013] Map any actual inspection path onto the configuration distribution map to form a pixel path with directional order. Determine the inspection route according to the order in which the pixel path overlaps with the field of view coverage map and the number and position of the overlapping pixels.

[0014] After correcting the camera control parameters, the final inspection route and configuration distribution map are mapped and stored in the inspection configuration library.

[0015] In this invention, the configuration distribution map uses the actual inspection area's plan design as the background and sets a corresponding scale to assist in placing cameras. For cameras such as bullet cameras that have a fixed orientation, the orientation of the camera in the map needs to be adjusted to match the actual camera orientation. For determining the inspection route, the user can directly draw an inspection path on the configuration distribution map, or present the actual inspection path on the configuration distribution map, thus obtaining a pixel path composed of continuous pixels, which includes a specific path direction and sequence. The playback order is determined by generating a camera sequence list based on the order of the camera's field of view coverage area traversed by the pixel path. Furthermore, the more pixels a pixel path contains within the field of view coverage area of ​​a certain camera, the longer the video playback time for that camera.

[0016] Preferably, the process of obtaining the view coverage map in the configuration distribution map includes:

[0017] Install cameras on the floor plan of the area to be inspected, ensuring their actual location and orientation match the cameras' orientation.

[0018] Set the camera's field of view threshold and convert it into pixel distance C on the configuration distribution map. iR , i represents the i-th camera; create a matrix Mask with the same size as the configuration distribution map, and select camera C. i Within the field of view of camera C i The distance is less than or equal to C iR All pixels p(x,y) are marked as Mask(p(x,y)) = C i ;

[0019] Traverse all cameras C i C i If all values ​​are greater than zero and are distinct, a field-of-view coverage map of the camera is obtained.

[0020] In this invention, the field of view coverage map is a superimposed combination of the field of view coverage of each camera. Some areas are simultaneously within the field of view coverage of multiple cameras, while some areas are in the camera's blind spot. The field of view distance threshold for each camera can be uniformly set to a default value first, and then adjusted for different cameras according to the actual situation. The field of view distance threshold is the farthest distance that the camera can monitor when projected on the configuration distribution map, and the field of view range is the maximum angle range that the camera can monitor when projected on the configuration distribution map.

[0021] Preferably, the process of determining the inspection route includes:

[0022] For a pixel path L, iterate through all pixels pl(x,y) on L. If a pixel pl(x,y) is covered by the field of view coverage map, add the camera with the corresponding field of view coverage to the camera order list.

[0023] Calculation with camera C i The shooting angle is obtained by using the direction vector from the starting point to the pixel pl(x,y). When several consecutive pixels have the same shooting angle relative to the same camera, the playback time is extended and stored in the camera control parameters; the playback time is the same for each pixel.

[0024] In this invention, the inspection route is determined based on the order in which the pixel path coincides with the field of view coverage map, as well as the number and position of the overlapping pixels. The coincidence of the pixel path with the field of view coverage map indicates that the inspection path passes through the field of view coverage of the corresponding camera. The camera order list is then arranged according to the order in which the inspection path passes through the camera's field of view coverage. When the path passes through the field of view coverage of multiple cameras simultaneously, the corresponding multiple cameras are added to the camera order list in parallel, and the video is played simultaneously. Each pixel in the pixel path corresponds to the same playback time. Therefore, the more the pixel path coincides with the camera's field of view coverage, the longer the video playback time for the corresponding camera.

[0025] As a preferred option, for camera C with a 360-degree field of view i With camera C i Centered on the camera, the system iterates through the surrounding pixels using a breadth-first search algorithm, calculating the distance from pixel p(x,y) to camera C. i distance d p ;

[0026] If d p ≤C iR If the pixel is not a wall point, then it is marked as Mask(p(x,y)) = C i Add a view coverage map;

[0027] If d p >C iR If a pixel is a wall point, skip to the next pixel check.

[0028] In this invention, cameras with a 360-degree field of view include PTZ cameras and bullet cameras with pan-tilt units. During calculations, the camera is treated as a point mass, with p(x) ci ,y ci ) represents the pixel where the camera is located; at the same time, since there are walls or other obstructions that block the camera in the actual area, the corresponding pixel on the configuration distribution map is uniformly set as the wall point. Then, the corresponding camera field of view coverage is obtained by comprehensively considering the field of view distance threshold and the wall point limitation range.

[0029] As a preferred option, for camera C with a field of view of a cone-shaped area i With camera C i Centered on the camera, the system iterates through the surrounding pixels using a breadth-first search algorithm, calculating the distance from pixel p(x,y) to camera C. i distance d p Pixel p(x,y) and camera C i The angle θ between the line connecting the two cameras and the camera's orientation;

[0030] If d p ≤C iR and If the pixel is not a wall point, then it is marked as Mask(p(x,y)) = C i Add a view coverage map;

[0031] If d p >C iR or If a pixel is a wall point, skip to the next pixel check.

[0032] In this invention, the camera with a field of view covering a conical area includes a bullet camera, etc., and the included angle of its field of view on the configuration distribution map is... The two rays of the camera are fixed in direction, and the extension of their direction bisects the angle of the field of view; in the calculation, the camera is treated as a point mass, with p(x) ci ,y ci ) represents the pixel where the camera is located; at the same time, since there are walls or other obstructions that block the camera in the actual area, the corresponding pixel on the configuration distribution map is uniformly set as the wall point. Then, the corresponding camera field of view coverage range is obtained by comprehensively considering the field of view distance threshold, the field of view angle and the wall point limitation range.

[0033] Preferably, when the distance between pixel p(x,y) and several cameras is less than the corresponding field of view distance threshold, all cameras that meet the conditions are marked on pixel p(x,y).

[0034] In this invention, for pixels that are simultaneously within the field of view of multiple cameras, all cameras that meet the conditions are marked on the pixel. When the real-time video of the cameras is played in the order of the inspection route, when a pixel is marked by multiple cameras, the real-time video of these cameras is played simultaneously, so as to monitor the actual area corresponding to the pixel in all directions.

[0035] Preferably, the camera control parameters include camera orientation, rotation speed, shooting angle, preset position setting, and preset position dwell time; the configuration distribution map is an SVG vector graphic, marked with walls and cameras, and the camera positions and orientations are consistent with those of real cameras.

[0036] In this invention, the camera orientation includes the fixed orientation of the camera, the rotation speed is the angular velocity of the camera when the shooting angle changes, the shooting angle is the angle between the line connecting the target pixel and the camera and the camera orientation, and the preset position setting refers to the camera position and shooting angle corresponding to some key monitoring areas; the configuration distribution map adopts SVG vector graphics. By utilizing the scalability and interactivity of SVG, a large-size distribution map can be drawn on the display interface, which can clearly and intuitively show the deployment and display of cameras and inspection routes. At the same time, it is saved in SVG format, which is conducive to subsequent local modification and customization.

[0037] This invention has the following beneficial effects: By setting the camera sequence list and camera control parameters in the inspection route, it fully utilizes camera resources for video inspection playback. Simultaneously, the configuration distribution map intuitively displays the camera locations, monitoring areas, and inspection routes, and stores the corresponding configuration distribution map and inspection routes in the inspection configuration library for easy retrieval and modification later. The configuration distribution map uses SVG vector graphics, leveraging the scalability and interactivity of SVG to clearly and intuitively enlarge and display the cameras and inspection routes. This invention can preset and store multiple inspection routes in the inspection configuration library, facilitating direct selection during subsequent inspections. Attached Figure Description

[0038] Figure 1 This is a flowchart of the video inspection method of the present invention;

[0039] Figure 2 This is a schematic diagram of the camera positions and orientations in the configuration distribution map of this invention embodiment;

[0040] Figure 3 This is the field-of-view coverage map in the configuration distribution map of this invention embodiment;

[0041] Figure 4 This is a schematic diagram of matrix mask markings when the field of view coverage of multiple cameras overlaps in an embodiment of the present invention;

[0042] Figure 5 This is a schematic diagram showing the inspection route on the configuration distribution map in an embodiment of the present invention;

[0043] Figure 6 This is a schematic diagram of the camera sequence list and camera control parameters included in the inspection route in this embodiment of the invention. Detailed Implementation

[0044] The present invention will now be further described with reference to the accompanying drawings and specific embodiments.

[0045] like Figure 1 As shown, a video inspection method based on 2D visualization includes:

[0046] Retrieve the inspection configuration library, parse and display the configuration distribution map and inspection route in the inspection configuration library;

[0047] Select the inspection route corresponding to the configuration distribution map to perform video inspection. The inspection route includes the camera sequence list and camera control parameters.

[0048] The system retrieves and plays the live video from the corresponding camera according to the camera sequence list, and controls the video capture of the camera playing the video according to the camera control parameters. When it is necessary to multicast the live video from multiple cameras, several windows pop up, each window playing the live video from one camera.

[0049] The process of building the inspection configuration library includes:

[0050] Obtain the configuration distribution map of the area to be inspected, including the location, orientation, and field of view coverage of the cameras;

[0051] Map any actual inspection path onto the configuration distribution map to form a pixel path with directional order. Determine the inspection route according to the order in which the pixel path overlaps with the field of view coverage map and the number and position of the overlapping pixels.

[0052] After correcting the camera control parameters, the final inspection route and configuration distribution map are mapped and stored in the inspection configuration library.

[0053] This invention primarily addresses situations involving large inspection areas, numerous inspection routes, and high inspection frequencies. It binds relevant information from actual cameras to corresponding cameras on a configuration distribution map. Users can draw inspection routes on the map, which automatically include all surrounding cameras, forming an initial camera sequence list and corresponding camera control parameters. Users can modify the inspection routes, including camera rotation speed and viewing angle dwell time. After configuration, the configuration distribution map and inspection routes are stored in the inspection configuration library for later retrieval. Users can choose to play a customized inspection route. Through a video playback component, the actual cameras are retrieved according to the configuration distribution map and inspection route, and real-time video is displayed on the playback interface, enabling online inspection tasks.

[0054] In this invention, the configuration distribution map uses the actual inspection area's plan design as the background and sets a corresponding scale to assist in placing cameras. For cameras such as bullet cameras that have a fixed orientation, the orientation of the camera in the map needs to be adjusted to match the actual camera orientation. For determining the inspection route, the user can directly draw an inspection path on the configuration distribution map, or present the actual inspection path on the configuration distribution map, thus obtaining a pixel path composed of continuous pixels, which includes a specific path direction and sequence. The playback order is determined by generating a camera sequence list based on the order of the camera's field of view coverage area traversed by the pixel path. Furthermore, the more pixels a pixel path contains within the field of view coverage area of ​​a certain camera, the longer the video playback time for that camera.

[0055] In this invention, the field of view coverage map is a superimposed combination of the field of view coverage of each camera. Some areas are simultaneously within the field of view coverage of multiple cameras, while some areas are in the camera's blind spot. The field of view distance threshold for each camera can be uniformly set to a default value first, and then adjusted for different cameras according to the actual situation. The field of view distance threshold is the farthest distance that the camera can monitor when projected on the configuration distribution map, and the field of view range is the maximum angle range that the camera can monitor when projected on the configuration distribution map.

[0056] In this invention, the inspection route is determined based on the order in which the pixel path coincides with the field of view coverage map, as well as the number and position of the overlapping pixels. The coincidence of the pixel path with the field of view coverage map indicates that the inspection path passes through the field of view coverage of the corresponding camera. The camera order list is then arranged according to the order in which the inspection path passes through the camera's field of view coverage. When the path passes through the field of view coverage of multiple cameras simultaneously, the corresponding multiple cameras are added to the camera order list in parallel, and the video is played simultaneously. Each pixel in the pixel path corresponds to the same playback time. Therefore, the more the pixel path coincides with the camera's field of view coverage, the longer the video playback time for the corresponding camera.

[0057] In this invention, cameras with a 360-degree field of view include PTZ cameras and bullet cameras with pan-tilt units. During calculations, the camera is treated as a point mass, with p(x) ci ,y ci ) represents the pixel where the camera is located; at the same time, since there are walls or other obstructions that block the camera in the actual area, the corresponding pixel on the configuration distribution map is uniformly set as the wall point. Then, the corresponding camera field of view coverage is obtained by comprehensively considering the field of view distance threshold and the wall point limitation range.

[0058] In this invention, the camera with a field of view covering a conical area includes a bullet camera, etc., and the included angle of its field of view on the configuration distribution map is... The two rays of the camera are fixed in direction, and the extension of their direction bisects the angle of the field of view; in the calculation, the camera is treated as a point mass, with p(x) ci ,y ci ) represents the pixel where the camera is located; at the same time, since there are walls or other obstructions that block the camera in the actual area, the corresponding pixel on the configuration distribution map is uniformly set as the wall point. Then, the corresponding camera field of view coverage range is obtained by comprehensively considering the field of view distance threshold, the field of view angle and the wall point limitation range.

[0059] In this invention, for pixels that are simultaneously within the field of view of multiple cameras, all cameras that meet the conditions are marked on the pixel. When the real-time video of the cameras is played in the order of the inspection route, when a pixel is marked by multiple cameras, the real-time video of these cameras is played simultaneously, so as to monitor the actual area corresponding to the pixel in all directions.

[0060] In this invention, the camera orientation includes the fixed orientation of the camera, the rotation speed is the angular velocity of the camera when the shooting angle changes, the shooting angle is the angle between the line connecting the target pixel and the camera and the camera orientation, and the preset position setting refers to the camera position and shooting angle corresponding to some key monitoring areas; the configuration distribution map adopts SVG vector graphics. By utilizing the scalability and interactivity of SVG, a large-size distribution map can be drawn on the display interface, which can clearly and intuitively show the deployment and display of cameras and inspection routes. At the same time, it is saved in SVG format, which is conducive to subsequent local modification and customization.

[0061] Camera preset positions refer to the way the key areas to be monitored are linked to the operating status of the camera. When the camera rotates to the area that needs to be monitored, a command to set a preset point is sent to the camera. The camera records the orientation angle and status at this time and associates it with the number of the preset point. When a recall command is sent to the camera, it will immediately move to the preset point and return to the state it was in at that time, so that the monitoring personnel can quickly view the monitored area.

[0062] In an embodiment of the present invention, the process of obtaining the view coverage map in the configuration distribution map includes:

[0063] Install cameras on the floor plan of the area to be inspected, ensuring their actual location and orientation match the cameras' orientation; for example... Figure 2As shown, a floor plan of the area to be inspected is imported into the SVG area as the background, and a scale is set to assist in placing the camera model. User operations are obtained through JavaScript. First, the walls in the picture are marked, and then the camera is dragged onto the floor plan. If it is a bullet camera or other camera with a fixed orientation, the orientation of the model needs to be adjusted to match the orientation of the actual camera, thereby generating a configuration distribution map that matches the actual camera distribution.

[0064] Set the camera's field of view threshold and convert it into pixel distance C on the configuration distribution map. iR Let i represent the i-th camera. In the initial state, the field of view distance threshold of all cameras can be uniformly set to a fixed value, such as 15m, and converted into pixel distance on the configuration distribution map using the corresponding scale. At the same time, the field of view distance threshold of the cameras can be manually modified according to the actual situation, such as increasing the field of view distance threshold in open areas and decreasing the field of view distance threshold in dimly lit spaces.

[0065] Create a matrix mask with the same dimensions as the configuration distribution map, and select camera C. i Within the field of view of camera C i The distance is less than or equal to C iR All pixels p(x,y) are marked as Mask(p(x,y)) = C i .

[0066] For camera C with a 360-degree field of view i Such as PTZ cameras or bullet cameras with pan-tilt units, using camera C i The pixel p(x) ci ,y ci Centered on the camera, the system iterates through the surrounding pixels using a breadth-first search algorithm, calculating the distance from pixel p(x,y) to camera C. i distance d p ;

[0067] If d p ≤C iR If the pixel is not a wall point, then it is marked as Mask(p(x,y)) = C i This pixel is added to the field-of-view coverage map;

[0068] If d p >C iR If a pixel is a wall point, then skip that pixel and proceed to the next pixel.

[0069] For camera C with a field of view of a cone-shaped area i For example, the angle of view of a bolt carrier on a deployment map is... With camera C iCentered on the camera, the system iterates through the surrounding pixels using a breadth-first search algorithm, calculating the distance from pixel p(x,y) to camera C. i distance d p Pixel p(x,y) and camera C i The angle θ between the line connecting the two cameras and the camera's orientation;

[0070] If d p ≤C iR and If the pixel is not a wall point, then it is marked as Mask(p(x,y)) = C i If d p >C iR or If a pixel is a wall point, skip to the next pixel check.

[0071] When the distance between pixel p(x,y) and several cameras is less than the corresponding field-of-view distance threshold, all cameras that meet the condition are marked on pixel p(x,y). Figure 4 As shown, when camera 1 is a PTZ camera and camera 2 is a bullet camera, there is a partial overlap in the field of view coverage of the PTZ camera and the bullet camera. For this part of the pixel p(x,y), it is marked as Mask(p(x,y)) = camera 1 and camera 2. When the pixel is only within the field of view coverage of one camera, only the corresponding camera is marked. For the area in the figure where Mask(p(x,y)) ! = null, it represents the field of view coverage of the camera, and the area where Mask(p(x,y)) == null represents the area not covered by the camera.

[0072] Traverse all cameras C i C i If all values ​​are greater than zero and are distinct, then the following can be obtained: Figure 3 The image shows the field of view coverage of the camera.

[0073] The process of determining the inspection route includes:

[0074] like Figure 5 As shown, any actual inspection path is mapped onto the configuration distribution map to form a pixel path with directional order; or a pixel path with directional order is drawn directly on the configuration distribution map. The method for drawing the path is based on a JavaScript file, which includes action logic, external interfaces, and SVG DOM objects.

[0075] For a pixel path L, iterate through all pixels pl(x,y) on L. If Mask(pl(x,y)) != null, it means that pixel pl(x,y) is covered by the field of view coverage map. Add the cameras within the corresponding field of view coverage area to the camera order list according to the pixel path order. The number of pixels that pixel path L passes through the field of view coverage area of ​​each camera is different. The total playback time of the camera video is determined based on the number of pixels in L within the field of view coverage area of ​​each camera. When pixel pl(x,y) marks only one camera, it means that the carousel will only play the real-time video of the corresponding camera. When pixel pl(x,y) marks multiple cameras, it means that the multicast will play the real-time video of multiple cameras simultaneously.

[0076] For cameras with a 360-degree field of view, such as PTZ cameras or bullet cameras with pan-tilt units, the calculation is based on camera C. i The shooting angle is obtained by using the direction vector from the starting point to the pixel pl(x,y). When several consecutive pixels have the same shooting angle relative to the same camera, the playback time is extended and the camera control parameters are stored. When the shooting angle changes, the camera is rotated to the corresponding angle and the camera sequence list data is added, while the corresponding camera control parameters are stored. The playback time is the same for each pixel.

[0077] The final result is obtained through the above steps. Figure 6 The inspection route shown includes a list of camera sequences and camera control parameters that can be displayed on the interface. The camera control parameters include information about all cameras in the area corresponding to the monitored path pixels, such as IP address and unique number to identify the camera; it also includes camera orientation (the default orientation is the orientation that is consistent with the actual camera on the configuration distribution map, and the target orientation of the corresponding path pixels), rotation speed, shooting angle, preset position setting, and preset position dwell time; the configuration distribution map is an SVG vector map, marked with walls and cameras, and the camera positions and orientations are consistent with the actual cameras.

[0078] After obtaining the configuration distribution map and inspection routes, the map and routes are stored in the server's inspection configuration library in JSON format. This facilitates subsequent retrieval and display of the configuration distribution map and inspection routes on the interface.

[0079] In this embodiment, the user directly retrieves the configuration distribution map and inspection routes from the inspection configuration library on the display interface, and selects one of the inspection routes to start video inspection. The system parses JSON data to obtain the camera order list and camera control parameters. Based on the camera order list, it sequentially retrieves and displays the real-time video from the corresponding cameras, while simultaneously controlling the operation of the camera playing the video according to the camera control parameters. In the camera order list, when playing a single camera's real-time video in a carousel, only one video window is displayed; when playing multiple cameras' real-time videos simultaneously in a multicast, multiple windows pop up, which can be arranged in an N*M format: 2 cameras in a 1*2 format, 3-4 cameras in a 2*2 format, 5-6 cameras in a 2*3 format, 7-9 cameras in a 3*3 format, and so on. Each window corresponds to playing the real-time video from one camera. The display interface can also optionally display the configuration distribution map and inspection routes, making it easier for the user to understand the specific monitoring area and the paths monitored by the cameras.

[0080] The above embodiments are further elaborations and descriptions of the present invention to facilitate understanding, and are not intended to limit the present invention in any way. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A video inspection method based on 2D visualization, characterized in that, include: Retrieve the inspection configuration library, parse and display the configuration distribution map and inspection route in the inspection configuration library; The configuration distribution map includes the location, orientation, and field of view coverage of the cameras; Install cameras on the floor plan of the inspection area that match their actual location and orientation; Set the camera's field of view threshold and convert it into pixel distance on the configuration distribution map. ; Create a matrix mask with the same dimensions as the configuration distribution map, and select the cameras. Within the field of vision and The distance is less than or equal to pixels All marked as ; Traverse all Obtain the field-of-view coverage map of the cameras, where i represents the i-th camera; Select the inspection route corresponding to the configuration distribution map to perform video inspection. The inspection route includes the camera sequence list and camera control parameters. The system retrieves and plays the real-time video from the corresponding camera according to the camera sequence list, and controls the camera's video acquisition according to the camera control parameters. When it is necessary to multicast live video from multiple cameras, several windows will pop up, with each window corresponding to the live video from one camera.

2. The video inspection method based on 2D visualization according to claim 1, characterized in that, The process of building the inspection configuration library includes: Obtain the configuration distribution map of the area to be inspected, including the location, orientation, and field of view coverage of the cameras; Map any actual inspection path onto the configuration distribution map to form a pixel path with directional order. Determine the inspection route according to the order in which the pixel path overlaps with the field of view coverage map and the number and position of the overlapping pixels. After correcting the camera control parameters, the final inspection route and configuration distribution map are mapped and stored in the inspection configuration library.

3. The video inspection method based on 2D visualization according to claim 1, characterized in that, The process of determining the inspection route includes: For a pixel path L, iterate through all pixels on L. If pixel If a camera is covered by the field of view coverage map, the camera within the corresponding field of view coverage area will be added to the camera order list. Calculation with camera Pointing to the pixel from the starting point The shooting angle is obtained from the direction vector. When several consecutive pixels have the same shooting angle relative to the same camera, the playback time is extended and stored in the camera control parameters; the playback time is the same for each pixel.

4. The video inspection method based on 2D visualization according to claim 1, characterized in that, For cameras with a 360-degree field of view With camera Centered on the data, the system iterates through the surrounding pixels using a breadth-first search algorithm to calculate the pixel values. to camera distance ; like If the pixel is not a wall point, then it is marked as Add a view coverage map; like If a pixel is a wall point, skip to the next pixel check.

5. The video inspection method based on 2D visualization according to claim 1, characterized in that, For cameras with a field of view of a cone-shaped area Its field of view angle on the configuration distribution map is 2. With camera Centered on the data, the system iterates through the surrounding pixels using a breadth-first search algorithm to calculate the pixel values. to camera distance Pixel With camera The angle between the line connecting the two cameras and the camera's orientation ; like and If the pixel is not a wall point, then it is marked as Add a view coverage map; like or If a pixel is a wall point, skip to the next pixel check.

6. The video inspection method based on 2D visualization according to claim 1, characterized in that, When pixel When the distance to several cameras is less than the corresponding field-of-view distance threshold, all cameras that meet the condition are marked on a pixel. superior.

7. A video inspection method based on 2D visualization according to claim 2 or 3, characterized in that, The camera control parameters include camera orientation, rotation speed, shooting angle, preset position setting, and preset position dwell time; the configuration distribution map is an SVG vector graphic, marked with walls and cameras, and the camera positions and orientations are consistent with those of real cameras.

Images