Power distribution switch cabinet state inspection method and system

Abstract

The invention provides a power distribution switch cabinet state inspection method, which comprises the following steps that: an inspection robot controls a camera to focus on a to-be-inspected surface for fixed-point video shooting, and a telescopic rod piece controls the camera to vertically move to realize fixed-point to whole video acquisition of the to-be-inspected surface, and further transmits the video to an industrial personal computer; the industrial personal computer edits the video to obtain a plurality of sub-region images containing different monitoring elements, the sub-region images are spliced into an overall image containing all the monitoring elements in the overall corridor, and position information of all the monitoring elements in the overall image is further recognized through an image recognition model; and identifying the state information of the monitoring elements in each sub-region image through a feature identification model, and performing image fusion on the whole image according to the position information and the state information of each monitoring element to obtain an inspection result image containing the state information of all the monitoring elements. According to the invention, the path of the inspection robot is optimized, the recognition efficiency is improved, and the state information of each monitoring element in the power distribution switch cabinet is clearly and visually previewed.

Description

Technical Field

[0001] The present invention relates to the technical field of power detection, and in particular to a method and system for inspecting the status of a power distribution switch cabinet. Background Art

[0002] Currently, inspection robots are widely used in the inspection of distribution switch cabinets. They can obtain the status information of the distribution switch cabinet in a timely manner to improve the convenience of inspection. Since the inspection robot needs to move to all monitoring components (such as pointers, pressure plates, digital meters, etc.) in the distribution switch cabinet in sequence according to the predetermined operation trajectory, and locate, shoot and identify according to the corresponding shooting positions, complex algorithms are required to design a complete set of operation trajectories and shooting positioning for the inspection robot, resulting in high costs and long time, and the recognition efficiency needs to be improved. At the same time, considering that there are many monitoring components involved in the distribution switch cabinet, the status information of the monitoring components is displayed in a messy manner, and it is impossible to preview them at a glance.

[0003] Therefore, in order to solve the problems of cumbersome operation trajectory and shooting positioning settings of the above-mentioned inspection robots during the inspection process, as well as the inconvenience in displaying the status information of the monitoring components, it is necessary to provide a new inspection method that can not only optimize the inspection robot path and improve the recognition efficiency, but also clearly and intuitively preview the status information of each monitoring component in the distribution switch cabinet. Summary of the Invention

[0004] The technical problem to be solved by the embodiments of the present invention is to provide a distribution switch cabinet status inspection method and system, which can not only optimize the inspection robot path and improve the recognition efficiency, but also clearly and intuitively preview the status information of each monitoring component in the distribution switch cabinet.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a method for inspecting the status of a distribution switch cabinet, the method comprising the following steps:

[0006] The inspection robot controls the camera mounted on it to focus on the inspection surface of the target distribution switch cabinet to shoot a video of a fixed area, and controls the camera to move in the vertical direction of the target distribution switch cabinet by extending and retracting the preset rods to complete the video acquisition from the fixed area to the entire area of the inspection surface, and further transmits the acquired video to the industrial computer;

[0007] The industrial control computer clips the videos collected by the inspection robot to obtain multiple sub-region images each containing at least one monitoring element and with all the monitoring elements contained being different. After the obtained multiple sub-region images are stitched into an overall image with the overall contour of the surface to be inspected and containing all the monitoring elements, the stitched overall image is further passed through a trained image recognition model to identify the position information of each monitoring element, and each sub-region image is passed through a preset plurality of feature recognition models to identify the status information of each monitoring element, and based on the position information of each monitoring element, the status information of each monitoring element is image-fused on the stitched overall image to obtain an inspection result image containing the status information of all the monitoring elements.

[0008] Among them, the specific steps for the inspection robot to control the camera mounted thereon to focus on the surface to be inspected of the target distribution switchgear for fixed-point area video shooting and to control the camera to move along the vertical direction of the target distribution switchgear by extending and retracting the preset rod to complete the video acquisition of the surface to be inspected from the fixed-point area to the overall area include:

[0009] The inspection robot controls the camera mounted thereon to focus on the surface to be inspected of the target distribution switchgear and performs fixed-point video acquisition on the top area of the surface to be inspected;

[0010] The inspection robot controls the rod to extend and retract along the vertical direction of the target distribution switchgear from top to bottom at a predetermined speed to drive the camera to perform video acquisition on the top area to the bottom area of the surface to be inspected, so as to complete the video acquisition of the surface to be inspected from the fixed-point area to the overall area.

[0011] Among them, the specific steps for the industrial control computer to clip the videos collected by the inspection robot to obtain multiple sub-region images each containing at least one monitoring element and with all the monitoring elements contained being different include:

[0012] The industrial control computer clips the videos collected by the inspection robot in units of frames to obtain multiple single-frame images arranged in consecutive frame numbers, extracts the single-frame images with the frame numbers being multiples of N and containing monitoring elements in the images, and compares the monitoring elements on all the extracted single-frame images to retain the single-frame images with all the monitoring elements contained being different and with the total number of monitoring elements in the retained single-frame images being all the monitoring elements on the surface to be inspected, and further outputs all the retained single-frame images as sub-region images; where 2 < N < 1 / 2 of the maximum frame number and N is a positive integer.

[0013] The industrial computer stitches the obtained multiple sub-region images into an overall image having the overall outline of the surface to be inspected and containing all monitoring elements, then passes the stitched overall image through a trained image recognition model to identify the position information of each monitoring element, and passes each sub-region image through a plurality of preset feature recognition models to identify the status information of each monitoring element, and performs image fusion on the status information of each monitoring element on the stitched overall image based on the position information of each monitoring element to obtain an inspection result image containing the status information of all monitoring elements. The specific steps include:

[0014] Based on the preset overall contour image of the surface to be inspected, the multiple sub-region images obtained are spliced together to obtain an overall image with the overall contour of the surface to be inspected and containing all monitoring elements;

[0015] Import the whole image into the trained image recognition model to identify the location information of each monitoring element;

[0016] Based on each sub-region image, a corresponding feature recognition model is selected from a plurality of preset feature recognition models for recognition to obtain status information of each monitoring element;

[0017] On the overall image, the status information of each monitoring element is fused according to its respective position information to obtain an inspection result image containing the status information of all monitoring elements.

[0018] Wherein, the monitoring elements include an instrument pointer, a hard pressure plate, an indicator light and a digital meter.

[0019] Wherein, the method further comprises:

[0020] The industrial computer performs three-dimensional modeling based on the physical structure of the target distribution switch cabinet, and combines the obtained inspection result image to obtain a three-dimensional model containing status information of all monitoring components.

[0021] Wherein, the method further comprises:

[0022] The industrial computer sends the obtained inspection result image or three-dimensional model containing status information of all monitoring components to the cloud.

[0023] The embodiment of the present invention further provides a distribution switch cabinet status inspection system, comprising an inspection robot and an industrial computer that are communicatively interconnected; wherein,

[0024] The inspection robot is used to control the camera installed thereon to focus on the surface to be inspected of the target power distribution switchgear for video shooting in a fixed-point area, and control the camera to move along the vertical direction of the target power distribution switchgear by extending and retracting a preset rod, so as to complete the video acquisition of the surface to be inspected from the fixed-point area to the overall area, and further transmit the collected video to the industrial control computer;

[0025] The industrial control computer is used to edit the video collected by the inspection robot to obtain multiple sub-region images containing at least one monitoring element and all the monitoring elements contained therein are different. After the obtained multiple sub-region images are spliced into an overall image with the overall contour of the surface to be inspected and containing all the monitoring elements, further pass the spliced overall image through a trained image recognition model to identify the position information of each monitoring element, and pass each sub-region image through a preset multiple feature recognition models to identify the state information of each monitoring element, and perform image fusion on the state information of each monitoring element on the spliced overall image according to the position information of each monitoring element, so as to obtain an inspection result image containing the state information of all monitoring elements.

[0026] Among them, the inspection robot includes:

[0027] The initial fixed-point area acquisition unit is used to control the camera installed thereon to focus on the surface to be inspected of the target power distribution switchgear and perform fixed-point video acquisition on the top area of the surface to be inspected;

[0028] The overall area acquisition unit is used to control the rod to extend and retract along the vertical direction of the target power distribution switchgear from top to bottom at a predetermined speed, so as to drive the camera to perform video acquisition on the top area to the bottom area of the surface to be inspected, so as to complete the video acquisition of the surface to be inspected from the fixed-point area to the overall area;

[0029] The video upload unit is used to transmit the collected video to the industrial control computer.

[0030] Among them, the industrial control computer includes:

[0031] The image editing unit is used to edit the video collected by the inspection robot in units of frames to obtain multiple single-frame images arranged in consecutive frame numbers, extract the single-frame images with frame numbers that are multiples of N and contain monitoring elements in the images, and compare the monitoring elements on all the extracted single-frame images, so as to retain the single-frame images with different monitoring elements contained therein and make the total number of monitoring elements in the retained single-frame images be all the monitoring elements on the surface to be inspected, and further output all the retained single-frame images as sub-region images; where 2 < N < 1 / 2 of the maximum frame number, and N is a positive integer;

[0032] An image stitching unit is used to stitch the multiple sub-region images obtained based on a preset overall outline image of the surface to be inspected, so as to obtain an overall image having the overall outline of the surface to be inspected and containing all monitoring elements;

[0033] A monitoring position recognition unit is used to import the entire image into a trained image recognition model to identify the position information of each monitoring element;

[0034] A monitoring status recognition unit is used to select a corresponding feature recognition model from a plurality of preset feature recognition models based on each sub-region image to perform recognition so as to obtain status information of each monitoring element;

[0035] The monitoring status fusion unit is used to fuse the status information of each monitoring element according to their respective position information on the overall image to obtain an inspection result image containing the status information of all monitoring elements.

[0036] The implementation of the embodiments of the present invention has the following beneficial effects:

[0037] 1. The inspection robot of the present invention controls the camera to move in the vertical direction of the target distribution switchgear by telescoping the preset rods, so as to complete the video acquisition from the fixed area to the entire area of the inspection surface. This eliminates the need to adjust the inspection robot to different fixed shooting positions to acquire video of different monitoring components, thus optimizing the inspection robot path and improving recognition efficiency.

[0038] 2. In the present invention, the industrial computer edits the video to splice out an overall image containing all monitoring elements, and identifies the location information and status information of each monitoring element through a trained image recognition model and multiple feature recognition models, and performs image fusion on the overall image to obtain an inspection result image containing the status information of all monitoring elements. This allows a clear and intuitive preview of the status information of each monitoring element in the distribution switch cabinet, without the need to frequently retrieve the status information images of different monitoring elements, saving time and reducing the error rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, without paying any creative work, other drawings obtained based on these drawings still fall within the scope of the present invention.

[0040] Figure 1 A flow chart of a method for inspecting the status of a power distribution switch cabinet provided by an embodiment of the present invention;

[0041] Figure 2A schematic structural diagram of a power distribution switch cabinet status inspection system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be described in further detail below with reference to the accompanying drawings.

[0043] like Figure 1 As shown in FIG. 1 , a method for inspecting the status of a power distribution switch cabinet is provided in an embodiment of the present invention. The method is implemented on an inspection robot and an industrial computer that are interconnected in communication. The method includes the following steps:

[0044] Step S1: The inspection robot controls the camera mounted thereon to focus on the inspection surface of the target distribution switch cabinet to capture a video of a fixed area, and controls the camera to move in the vertical direction of the target distribution switch cabinet by extending and retracting a preset rod to complete video capture of the inspection surface from the fixed area to the entire area, and further transmits the captured video to the industrial computer;

[0045] The specific process is as follows: first, the inspection robot arrives at the target distribution switchgear according to a preset motion trajectory. Then, according to an inspection instruction, it controls its mounted camera (not shown) to focus on the target distribution switchgear (not shown) to inspect the surface to be inspected (e.g., the front of all electrical equipment installed) and capture a fixed-point video. It should be noted that the inspection robot's motion process and method of reaching the target distribution switchgear are implemented using common techniques in the field and will not be detailed here.

[0046] In one example, the inspection robot controls the camera's field of view so that the left and right sides of the target distribution switch cabinet appear completely and clearly in the camera's field of view. The camera is further directed at the top area of the surface to be inspected (such as the front of the target distribution switch cabinet), and the camera is turned on to perform fixed-point video capture of the top area of the surface to be inspected.

[0047] Secondly, according to another inspection instruction, the inspection robot controls the camera to move in the vertical direction of the target distribution switch cabinet by extending and retracting the preset rod (not shown) to complete the video acquisition of the inspection surface from the fixed area to the entire area.

[0048] In one example, the inspection robot controls the telescopic rod to move downward along the vertical direction of the target distribution switchgear at a predetermined speed (e.g., 3 m / s), so that the camera sequentially records the video of the surface to be inspected (such as the front of the target distribution switchgear) along the vertical direction from the top area of the target distribution switchgear until the bottom of the target distribution switchgear appears in the camera's field of view. There is no jitter in the process, that is, the camera is driven to collect the video of the surface to be inspected (such as the front of the target distribution switchgear) from the top area to the bottom area to complete the video collection of the surface to be inspected (such as the front of the target distribution switchgear) from the fixed-point area to the overall area. It should be noted that the movement direction of the telescopic rod can be flexibly adjusted according to the actual situation, but it is necessary to ensure that the camera can complete the video collection of the surface to be inspected from the fixed-point area to the overall area.

[0049] Finally, the inspection robot transmits the collected video to the industrial control computer.

[0050] Step S2: The industrial control computer clips the video collected by the inspection robot to obtain multiple sub-region images containing at least one monitoring component and all the monitoring components contained are different. After the obtained multiple sub-region images are spliced into an overall image with the overall contour of the surface to be inspected and containing all the monitoring components, the spliced overall image is further passed through the trained image recognition model to identify the position information of each monitoring component, and each sub-region image is passed through a preset multiple feature recognition models to identify the status information of each monitoring component, and according to the position information of each monitoring component, the status information of each monitoring component is image-fused on the spliced overall image to obtain an inspection result image containing the status information of all the monitoring components.

[0051] The specific process is as follows. First, the industrial control computer clips the video collected by the inspection robot in units of frames to obtain multiple single-frame images arranged in consecutive frame numbers, and extracts the single-frame images with frame numbers that are multiples of N (e.g., 4) and contain monitoring components (such as instrument pointers, hard pressure plates, indicator lights, or digital meters, etc.) in the images, and compares the monitoring components on all the extracted single-frame images to retain the single-frame images with all different monitoring components contained and make the total number of monitoring components in the retained single-frame images be all the monitoring components on the surface to be inspected. Further, all the retained single-frame images are output as sub-region images; where 2 < N < 1 / 2 of the maximum frame number, and N is a positive integer.

[0052] In one example, the industrial computer cuts 24 consecutive frames from a video and, based on the camera's movement speed, extracts single frames with frame numbers multiples of 4*, capturing a rectangular area in the center of the image containing the monitoring element. The captured area is ensured to be free of noticeable distortion. For example, the computer extracts single frames at frames 4, 8, 16, 20, and 24, ensuring that each of these five frames contains a different monitoring element.

[0053] The monitoring elements on these five frames are compared to retain single frames containing different monitoring elements, ensuring that the monitoring elements in the retained single frames cover all monitoring elements on the surface to be inspected. These retained single frames are then output as sub-region images. For example, if the monitoring elements in the 8th and 16th frames are identical, if the monitoring elements in the 8th frame and the other three frames (excluding the 8th and 16th frames) cover all monitoring elements on the surface to be inspected (e.g., the front of the target distribution switchgear), then the single frames of the 4th, 8th, 20th, and 24th frames are retained as sub-region images. Otherwise, if the monitoring elements in the 16th frame and the other three frames (excluding the 8th and 16th frames) cover all monitoring elements on the surface to be inspected (e.g., the front of the target distribution switchgear), then the single frames of the 4th, 16th, 20th, and 24th frames are retained as sub-region images.

[0054] Secondly, based on the preset overall contour image of the surface to be inspected, the obtained multiple sub-region images are spliced to obtain an overall image having the overall contour of the surface to be inspected and containing all monitoring elements.

[0055] In one example, based on the overall contour of the surface to be inspected, information of two adjacent sub-region images is matched in sequence to form an overall image of the surface to be inspected (such as the front of the target distribution switch cabinet).

[0056] Then, the overall image is imported into the trained image recognition model to identify the location information of each monitoring element.

[0057] In one example, an image recognition model is constructed in advance based on the YOLO neural network, and the image recognition model is trained by pre-labeling the image information of a large number of monitoring elements on the surface to be inspected (such as the front of the target distribution switch cabinet), thereby obtaining a trained image recognition model. The above-mentioned spliced overall image is then imported into the trained image recognition model to extract the position information of each monitoring element in the overall image.

[0058] Next, based on each sub-region image, a corresponding feature recognition model is selected from a plurality of preset feature recognition models for recognition, so as to obtain status information of each monitoring element.

[0059] In one example, a feature recognition model is pre-built for each monitoring element based on its type. Then, based on the type of monitoring element in each sub-region image, one or more corresponding feature recognition models are selected for status recognition, thereby obtaining status information for each monitoring element. This information includes, for example, the reading of an instrument pointer, the engagement and disengagement status of a hard pressure plate, indicator light status, and the reading of a digital meter.

[0060] Finally, on the overall image, the status information of each monitoring element is fused according to its respective position information to obtain an inspection result image containing the status information of all monitoring elements.

[0061] In an embodiment of the present invention, the industrial computer also integrates the inspection result image information into the three-dimensional model, so the method further includes: the industrial computer performs three-dimensional modeling based on the physical structure of the target distribution switch cabinet, and combines the obtained inspection result image to obtain a three-dimensional model containing the status information of all monitoring components.

[0062] In addition, the industrial computer also uploads the integrated three-dimensional model or inspection result image to the cloud for display, so the method further includes: the industrial computer sends the obtained inspection result image or three-dimensional model containing the status information of all monitoring components to the cloud.

[0063] like Figure 2 As shown in the figure, a distribution switch cabinet status inspection system is provided in an embodiment of the present invention, including an inspection robot 1 and an industrial computer 2 that are interconnected in communication; wherein,

[0064] The inspection robot 1 is used to control the camera mounted thereon to focus on the inspection surface of the target distribution switch cabinet to capture a video of a fixed area, and to control the camera to move in the vertical direction of the target distribution switch cabinet by extending and retracting a preset rod to complete video capture from the fixed area to the entire area of the inspection surface, and further transmit the captured video to the industrial computer 2;

[0065] The industrial computer 2 is used to edit the video collected by the inspection robot 1 to obtain multiple sub-area images containing at least one monitoring element and each of which contains different monitoring elements. After the multiple sub-area images are spliced into an overall image having the overall outline of the surface to be inspected and containing all monitoring elements, the spliced overall image is further passed through a trained image recognition model to identify the position information of each monitoring element, and each sub-area image is passed through a preset plurality of feature recognition models to identify the status information of each monitoring element. Based on the position information of each monitoring element, the status information of each monitoring element is fused on the spliced overall image to obtain an inspection result image containing the status information of all monitoring elements.

[0066] Among them, the inspection robot 1 includes:

[0067] An initial fixed-point area acquisition unit 11, which is used to control the camera provided thereon to focus on the surface to be inspected of the target power distribution switchgear, and perform fixed-point video acquisition on the top area of the surface to be inspected;

[0068] An overall area acquisition unit 12, which is used to control the telescopic movement of the rod along the vertical direction of the target power distribution switchgear from top to bottom at a predetermined speed, so as to drive the camera to perform video acquisition on the top area to the bottom area of the surface to be inspected, so as to complete the video acquisition of the surface to be inspected from the fixed-point area to the overall area;

[0069] A video upload unit 13, which is used to transmit the collected video to the industrial control computer.

[0070] Among them, the industrial control computer 2 includes:

[0071] An image editing unit 21, which is used to edit the video collected by the inspection robot in units of frames, obtain a plurality of single-frame images arranged in consecutive frame numbers, extract the single-frame images whose frame numbers are multiples of N and the single-frame images containing monitoring elements, and compare the monitoring elements on all the extracted single-frame images, so as to retain the single-frame images in which the monitoring elements contained in the images are all different and make the total number of monitoring elements in the retained single-frame images be all the monitoring elements on the surface to be inspected, and further output all the retained single-frame images as sub-region images; where 2 < N < 1 / 2 of the maximum frame number, and N is a positive integer;

[0072] An image stitching unit 22, which is used to stitch the obtained multiple sub-region images based on the preset overall contour image of the surface to be inspected, and obtain an overall image with the overall contour of the surface to be inspected and containing all monitoring elements;

[0073] A monitoring position recognition unit 23, which is used to import the overall image into a trained image recognition model to recognize the position information of each monitoring element;

[0074] A monitoring state recognition unit 24, which is used to select the corresponding feature recognition model from a preset plurality of feature recognition models based on each sub-region image for recognition, so as to obtain the state information of each monitoring element;

[0075] A monitoring state fusion unit 25, which is used to perform image fusion on the state information of each monitoring element according to their respective position information on the overall image, and obtain an inspection result image containing the state information of all monitoring elements.

[0076] Implementing the embodiments of the present invention has the following beneficial effects:

[0077] 1. The inspection robot of the present invention controls the camera to move in the vertical direction of the target distribution switchgear by telescoping the preset rods, so as to complete the video acquisition from the fixed area to the entire area of the inspection surface. This eliminates the need to adjust the inspection robot to different fixed shooting positions to acquire video of different monitoring components, thus optimizing the inspection robot path and improving recognition efficiency.

[0078] 2. In the present invention, the industrial computer edits the video to splice out an overall image containing all monitoring elements, and identifies the location information and status information of each monitoring element through a trained image recognition model and multiple feature recognition models, and performs image fusion on the overall image to obtain an inspection result image containing the status information of all monitoring elements. This allows a clear and intuitive preview of the status information of each monitoring element in the distribution switch cabinet, without the need to frequently retrieve the status information images of different monitoring elements, saving time and reducing the error rate.

[0079] It is worth noting that in the above system embodiment, the various system modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the various functional modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0080] Those skilled in the art will understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the program can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc.

[0081] The above disclosure is merely a preferred embodiment of the present invention and certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for inspecting the status of a power distribution switch cabinet, characterized in that: The method includes the following steps: The inspection robot controls the camera mounted thereon to focus on the surface to be inspected of the target power distribution switchgear for video shooting of a fixed-point area, and controls the camera to move along the vertical direction of the target power distribution switchgear by means of a telescopic pre-set rod, so as to complete the video acquisition of the surface to be inspected from the fixed-point area to the overall area, and further transmits the acquired video to an industrial control computer; The industrial control computer edits the video collected by the inspection robot to obtain multiple sub-region images each containing at least one monitoring element and all the monitoring elements contained therein being different. After the obtained multiple sub-region images are stitched into an overall image with the overall contour of the surface to be inspected and containing all the monitoring elements, the industrial control computer further passes the stitched overall image through a trained image recognition model to identify the position information of each monitoring element, and passes each sub-region image through a preset plurality of feature recognition models to identify the state information of each monitoring element, and performs image fusion on the state information of each monitoring element on the stitched overall image according to the position information of each monitoring element, so as to obtain an inspection result image containing the state information of all the monitoring elements.

2. The method for inspecting the status of a distribution switch cabinet according to claim 1, wherein: The specific steps for the inspection robot to control the camera mounted thereon to focus on the surface to be inspected of the target power distribution switchgear for video shooting of a fixed-point area and to control the camera to move along the vertical direction of the target power distribution switchgear by means of a telescopic pre-set rod so as to complete the video acquisition of the surface to be inspected from the fixed-point area to the overall area include: The inspection robot controls the camera mounted thereon to focus on the surface to be inspected of the target power distribution switchgear and performs fixed-point video acquisition on the top area of the surface to be inspected; 3. The method for inspecting the status of a distribution switch cabinet according to claim 1, wherein: The inspection robot controls the telescopic rod to move along the vertical direction of the target power distribution switchgear from top to bottom at a predetermined speed, so as to drive the camera to perform video acquisition on the top area to the bottom area of the surface to be inspected, so as to complete the video acquisition of the surface to be inspected from the fixed-point area to the overall area. The specific steps for the industrial control computer to edit the video collected by the inspection robot to obtain multiple sub-region images each containing at least one monitoring element and all the monitoring elements contained therein being different include: The industrial control computer edits the video collected by the inspection robot frame by frame to obtain multiple single-frame images arranged in consecutive frame numbers, extracts the single-frame images with frame numbers being multiples of N and containing monitoring elements in the images, and compares the monitoring elements on all the extracted single-frame images, so as to retain the single-frame images in which all the monitoring elements contained in the images are different and make the total number of monitoring elements in the retained single-frame images be all the monitoring elements on the surface to be inspected, and further outputs all the retained single-frame images as sub-region images; where 2 < N < 1 / 2 of the maximum frame number, and N is a positive integer.

4. The method for inspecting the status of a distribution switch cabinet according to claim 3, wherein: After the industrial computer stitches the obtained multiple sub-region images into an overall image having the overall outline of the surface to be inspected and containing all monitoring elements, the stitched overall image is further subjected to a trained image recognition model to identify the position information of each monitoring element, and each sub-region image is subjected to a plurality of preset feature recognition models to identify the status information of each monitoring element, and image fusion is performed on the status information of each monitoring element on the stitched overall image based on the position information of each monitoring element to obtain an inspection result image containing the status information of all monitoring elements. The specific steps include: Based on the preset overall contour image of the surface to be inspected, the multiple sub-region images obtained are spliced together to obtain an overall image with the overall contour of the surface to be inspected and containing all monitoring elements; Import the whole image into the trained image recognition model to identify the location information of each monitoring element; Based on each sub-region image, a corresponding feature recognition model is selected from a plurality of preset feature recognition models for recognition to obtain status information of each monitoring element; On the overall image, the status information of each monitoring element is fused according to its respective position information to obtain an inspection result image containing the status information of all monitoring elements.

5. The method for inspecting the status of a distribution switch cabinet according to claim 4, wherein: The monitoring elements include an instrument pointer, a hard pressure plate, an indicator light and a digital meter.

6. The method for inspecting the status of a distribution switch cabinet according to claim 1, wherein: The method further comprises: The industrial computer performs three-dimensional modeling based on the physical structure of the target distribution switch cabinet, and combines the obtained inspection result image to obtain a three-dimensional model containing status information of all monitoring components.

7. The method for inspecting the status of a distribution switch cabinet according to claim 6, wherein: The method further comprises: The industrial computer sends the obtained inspection result image or three-dimensional model containing status information of all monitoring components to the cloud.

8. A distribution switch cabinet status inspection system, characterized in that: Including inspection robots and industrial computers with communication interconnection; among them, The inspection robot is used to control the camera mounted thereon to focus on the surface to be inspected of the target distribution switch cabinet to perform video capture of a fixed area, and to control the camera to move in the vertical direction of the target distribution switch cabinet by retracting the preset rod to complete video capture of the surface to be inspected from the fixed area to the entire area, and further transmit the captured video to the industrial computer; The industrial computer is used to edit the video collected by the inspection robot to obtain multiple sub-area images containing at least one monitoring element and each containing different monitoring elements. After the multiple sub-area images are spliced into an overall image having the overall outline of the surface to be inspected and containing all monitoring elements, the spliced overall image is further used to identify the position information of each monitoring element through a trained image recognition model, and each sub-area image is used to identify the status information of each monitoring element through a preset multiple feature recognition models. Based on the position information of each monitoring element, the status information of each monitoring element is fused on the spliced overall image to obtain an inspection result image containing the status information of all monitoring elements.

9. The distribution switch cabinet status inspection system according to claim 8, characterized in that: The inspection machine includes: The initial fixed-point area acquisition unit is used to control the camera set thereon to focus on the surface to be inspected of the target power distribution switchgear, and perform fixed-point video acquisition on the top area of the surface to be inspected; The overall area acquisition unit is used to control the telescopic movement of the rod along the vertical direction of the target power distribution switchgear from top to bottom at a predetermined speed, so as to drive the camera to perform video acquisition on the top area to the bottom area of the surface to be inspected, so as to complete the video acquisition of the surface to be inspected from the fixed-point area to the overall area; The video upload unit is used to transmit the acquired video to the industrial control computer.

10. The distribution switch cabinet status inspection system according to claim 8, characterized in that: The industrial control computer includes: The image editing unit is used to edit the video collected by the inspection robot in units of frames, obtain multiple single-frame images arranged in consecutive frame numbers, extract the single-frame images whose frame numbers are multiples of N and contain monitoring components in the images, and compare the monitoring components on all the extracted single-frame images, so as to retain the single-frame images in which the monitoring components contained in the images are all different and make the monitoring components in the retained single-frame images total all the monitoring components on the surface to be inspected, and further output all the retained single-frame images as sub-region images; where 2 < N < 1 / 2 of the maximum frame number, and N is a positive integer; The image stitching unit is used to stitch the obtained multiple sub-region images based on the preset overall contour image of the surface to be inspected, and obtain an overall image with the overall contour of the surface to be inspected and containing all monitoring components; The monitoring position recognition unit is used to import the overall image into a trained image recognition model to recognize the position information of each monitoring component; The monitoring status recognition unit is used to select the corresponding feature recognition model from a preset multiple feature recognition models based on each sub-region image for recognition, so as to obtain the status information of each monitoring component; The monitoring status fusion unit is used to perform image fusion on the overall image for the status information of each monitoring component according to their respective position information, and obtain an inspection result image containing the status information of all monitoring components.

Images