Unmanned substation state monitoring system, method and device based on image recognition

By adopting image recognition technology and multi-camera monitoring system in unmanned substations, the problem that traditional monitoring equipment cannot monitor the wear of staff throughout the process is solved, real-time monitoring of safety helmets, work clothes and insulated gloves in the substation is realized, and safety is improved and monitoring efficiency.

CN118945288AInactive Publication Date: 2024-11-12NARI TECH CO LTD

Patent Information

Application Number
CN202410789972.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-19
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Traditional unmanned substation monitoring equipment cannot achieve full monitoring of staff, especially when wearing safety helmets, work clothes and insulated gloves, which can easily lead to negligence and visual fatigue, and abnormal situations cannot be detected in time.

Method used

The unmanned substation status monitoring system based on image recognition is adopted, and the panoramic monitoring mechanism and tracking monitoring mechanism are used to use multiple cameras for panoramic and tracking monitoring. Combined with image acquisition, processing and alarm modules, the wear status of people is judged and recorded in real time, and Real-time tracking of personnel is achieved through adjustable mechanisms.

Benefits of technology

The full monitoring of staff in the substation is achieved, blind spots are avoided, and the wearing of safety helmets, work clothes and insulated gloves is monitored in a timely manner, improving safety and efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118945288A_ABST
    Figure CN118945288A_ABST
Patent Text Reader

Abstract

The invention discloses an unmanned substation state monitoring system, method and device based on image recognition, and the system comprises a panoramic monitoring mechanism and a tracking monitoring mechanism, the panoramic monitoring mechanism comprises four first monitoring units, and the four first monitoring units are respectively arranged at four corners of an unmanned substation. The tracking and monitoring mechanism comprises a second camera and a second adjustable mechanism, the second camera is arranged above the whole monitoring area through the second adjustable mechanism, panoramic monitoring of the whole transformer substation is achieved through the arrangement of the panoramic monitoring mechanism, and the tracking and monitoring mechanism is matched for tracking and monitoring personnel targets. A blind area of panoramic monitoring is avoided, and whether a worker wears safety helmets, working clothes and insulating gloves or not according to requirements is monitored in the whole process; the panoramic monitoring mechanism obtains entrance information and movement information of personnel, and the tracking monitoring mechanism is combined to obtain personnel movement information of the personnel, so that the real-time position adjustment of the tracking monitoring mechanism is realized.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of substation monitoring, and in particular to an unmanned substation status monitoring system, method and device based on image recognition. Background Art

[0002] The monitoring equipment used in traditional unmanned substations transmits the monitoring video information of the unmanned substation directly to the monitoring terminal, and personnel are deployed to stare at the display screen at the monitoring terminal. Manual monitoring is prone to visual fatigue, and dozing off at night can easily lead to negligence, and abnormal situations cannot be discovered in time.

[0003] Currently, monitoring equipment in substations is installed in fixed positions. Video surveillance is set up at the entrance and exit of substations to monitor whether workers are wearing safety helmets / work clothes / insulating gloves as required. However, workers' behavior after entering the substation cannot be monitored all the time. Summary of the invention

[0004] The purpose of the present invention is to provide an unmanned substation status monitoring system, method and device based on image recognition to address the deficiencies of the prior art.

[0005] The technical solution of the present invention to solve the above problem is: an unmanned substation status monitoring device based on image recognition, including a panoramic monitoring mechanism and a tracking monitoring mechanism, wherein the panoramic monitoring mechanism includes four first monitoring units, which are respectively arranged at the four corners of the unmanned substation.

[0006] The tracking monitoring mechanism includes a second camera and a second adjustable mechanism. The second camera is arranged above the entire monitoring area through the second adjustable mechanism. The second adjustable mechanism includes an X-axis adjustment mechanism and a Y-axis adjustment mechanism.

[0007] Furthermore, the first monitoring unit includes a first column and a first camera, the first camera is installed on the column through a first adjustable mechanism, the first adjustable mechanism includes a ring gear, a ring track and a first mounting seat, the ring gear and the ring track are respectively connected to the column through a plurality of connecting rods, and the ring gear and the ring track are coaxially arranged, the first mounting seat includes a track connector, a first mounting plate and a first driving mechanism, the first driving mechanism includes a first driving motor and a first driving gear, the first driving motor is installed on the first mounting plate through a frame, the first driving gear is coaxially connected to the output shaft of the first driving motor, the first driving gear is meshed with the ring gear, the track connector is fixedly connected to the first mounting plate through a connecting rod, an arc-shaped slot is provided in the middle of the track connector, the track connector is slidably mounted on the ring track through the arc-shaped slot, and the first camera is connected to the lower end surface of the track connector.

[0008] Furthermore, the X-axis adjustment mechanism includes an X-axis adjustment screw, an X-axis guide rod and a gantry, the X-axis adjustment screw and the X-axis guide rod are respectively arranged on the front and rear sides of the monitoring area through fixed mounting seats, both ends of the X-axis adjustment screw are rotatably connected to the fixed mounting seat through bearings, one end of the X-axis adjustment screw is connected to the output shaft of the X-axis adjustment motor through a coupling, the X-axis guide rod is fixedly connected to the fixed mounting seat, and the two ends of the gantry are respectively provided with a first connecting member and a first sliding member, the first connecting member is screwed on the X-axis adjustment screw, and the first sliding member is slidably sleeved on the X-axis guide rod, the Y-axis adjustment mechanism is arranged on the bottom surface of the gantry, and the Y-axis adjustment mechanism includes The invention comprises a Y-axis adjusting screw, a Y-axis guide rod and a base plate, wherein the two ends of the Y-axis adjusting screw are rotatably connected to the two fixing seats through bearings respectively, the two ends of the Y-axis guide rod are fixedly connected to the two fixing seats, one end of the Y-axis adjusting screw is connected to the driven gear, the driven gear is meshed with the second driving gear, the second driving gear is connected to the output shaft of the Y-axis adjusting motor, a second connecting piece is provided on the base plate, a threaded hole matching with the Y-axis adjusting screw is provided on the second connecting piece, and a guide hole matching with the Y-axis guide rod is also provided, the second connecting piece is screwed on the screw through the threaded hole, and is slidably sleeved on the guide rod through the guide hole, and the second camera is fixedly connected to the lower end surface of the base plate.

[0009] An unmanned substation status monitoring system based on image recognition comprises the above-mentioned monitoring device, as well as an image acquisition module, an image processing module, an alarm module, and a controller. Four first monitoring units and a second camera are respectively connected to the image processing module via the image acquisition modules, the image processing module is connected to the controller, and the X-axis adjustment mechanism and the Y-axis adjustment mechanism are respectively connected to the controller.

[0010] An unmanned substation status monitoring method based on image recognition specifically comprises the following steps:

[0011] Step 1: Adjust the positions of the four first monitoring units at the four corners respectively to ensure that all the entrances and exits of the substation area are within the range of panoramic monitoring;

[0012] Step 2: The four first monitoring units collect panoramic monitoring videos of the scene and transmit them to the image acquisition module. The image acquisition module encodes and packages the panoramic monitoring videos into RTSP video streams and transmits them to the image processing module for image processing and storage.

[0013] Step 3: The image processing module decodes and analyzes the video stream to determine whether the person is not wearing a safety helmet, work clothes, or insulating gloves.

[0014] Step 4: When it is determined that the person is not wearing a safety helmet, work clothes, or insulating gloves, the video clip is saved, the video time period is recorded, and an alarm is sounded through the alarm module.

[0015] Step 5: The image processing module decodes and analyzes the video stream, determines the movement direction of the personnel, and transmits the entrance number of the personnel entering the substation and the personnel movement information to the controller. The controller first adjusts the second adjustable mechanism to move the second camera to the corresponding substation entrance, and adjusts the second adjustable mechanism in real time according to the personnel movement information.

[0016] Step 6: The second camera collects the tracking monitoring video and transmits it to the image acquisition module. The image acquisition module encodes and packages the tracking monitoring video into an RTSP video stream and transmits it to the image processing module for image processing.

[0017] Step 7, the image processing module decodes and analyzes the video stream to determine whether the person is not wearing a safety helmet, work clothes, or insulating gloves, and determines the direction of movement of the person, and then transmits the person's movement information to the controller, and the controller adjusts the second adjustable mechanism in real time to track the person.

[0018] Furthermore, the image processing module determines and analyzes the moving direction of the person, specifically including the following steps:

[0019] Step b1, decoding and segmenting the video into multiple video frames, and sorting the video frames by time to obtain orderly arranged video frames,

[0020] Step b2, performing image noise reduction preprocessing on each video frame;

[0021] Step b3: Segment the objects in the scene based on the mean shift algorithm;

[0022] Step b4, extracting the moving target using the background difference method;

[0023] Step b5, preferentially use the panoramic surveillance video of the first camera to analyze and determine the movement direction and position of the person, compare the current video frame with the video frame of the person at the entrance and exit positions, and obtain the movement information of the person. When the person leaves the current monitoring screen of the first camera, obtain the current time t1, and preferentially call up the panoramic surveillance videos of other first cameras to continue to analyze and determine the movement direction and position of the person, compare the current video frame with the video frame of the person at time t1, and obtain the movement information of the person. When all the monitoring screens of the first cameras cannot capture the person, obtain the current time t2, call up the tracking surveillance video of the second camera to continue to analyze and determine the movement direction and position of the person, compare the current video frame with the video frame of the previous second, and obtain the movement information of the person.

[0024] The present invention has the beneficial effects:

[0025] The present invention provides an unmanned substation status monitoring system, method and device based on image recognition, which realizes panoramic monitoring of the entire substation by setting up a panoramic monitoring mechanism, cooperates with a tracking monitoring mechanism to track and monitor personnel targets, avoids blind spots in panoramic monitoring, and monitors whether the staff wears safety helmets, work clothes and insulating gloves as required; the panoramic monitoring mechanism obtains the entrance information and movement information of the personnel, and combines with the tracking monitoring mechanism to obtain the personnel movement information of the personnel, so as to realize real-time position adjustment of the tracking monitoring mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A top view of the present invention;

[0027] Figure 2 is a top view of the first monitoring unit;

[0028] Figure 3 is a front view of the first monitoring unit;

[0029] Figure 4 is a structural schematic diagram of a first mounting base;

[0030] Figure 5 It is a structural schematic diagram of the Y-axis adjustment mechanism;

[0031] Figure 6 is a schematic structural diagram of a second connecting member;

[0032] In the figure: 1-first monitoring unit, 2-first column, 3-first camera, 4-annular gear, 5-annular track, 6-first mounting seat, 7-track connector, 8-first mounting plate, 9-first drive motor, 10-first drive gear, 11-arc slot, 12-pulley, 13-second camera, 14-X-axis adjusting screw, 15-X-axis guide rod, 16-gantry, 17-first connecting member, 18-first sliding member, 19-Y-axis adjusting screw, 20-Y-axis guide rod, 21-bottom plate, 22-second connecting member. DETAILED DESCRIPTION

[0033] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood by specific circumstances.

[0034] As shown in the figure, an unmanned substation status monitoring device based on image recognition includes a panoramic monitoring mechanism and a tracking monitoring mechanism. The panoramic monitoring mechanism includes four first monitoring units, which are respectively arranged at the four corners of the unmanned substation to perform panoramic monitoring of the entire substation. The first monitoring unit includes a first column and a first camera. The first camera is installed on the column through a first adjustable mechanism. The first adjustable mechanism includes a ring gear, a ring track and a first mounting seat. The ring gear and the ring track are respectively connected to the column through a plurality of connecting rods, and the ring gear and the ring track are coaxially arranged. The first mounting seat includes a track connector, a first mounting seat, and a first mounting seat. The first driving mechanism comprises a first driving motor and a first driving gear. The first driving motor is mounted on the first mounting plate through a frame. The first driving gear is coaxially connected to the output shaft of the first driving motor. The first driving gear is meshed with a ring gear. The track connector is fixedly connected to the first mounting plate through a connecting rod. An arc-shaped slot is provided in the middle of the track connector. The track connector is slidably sleeved on the ring track through the arc-shaped slot. An inner groove and an outer groove are provided on both sides of the ring track. Two sets of pulleys are provided on both sides of the arc-shaped slot. The two sets of pulleys are respectively rotatably arranged in the inner groove and the outer groove to improve the smoothness of sliding. The first camera is connected to the lower end surface of the track connector.

[0035] The tracking monitoring mechanism includes a second camera and a second adjustable mechanism, the second camera is arranged above the entire monitoring area through the second adjustable mechanism, the second adjustable mechanism includes an X-axis adjustment mechanism and a Y-axis adjustment mechanism, the X-axis adjustment mechanism includes an X-axis adjustment screw, an X-axis guide rod and a gantry, the X-axis adjustment screw and the X-axis guide rod are respectively arranged on the front and rear sides of the monitoring area through fixed mounting seats, the two ends of the X-axis adjustment screw are rotatably connected to the fixed mounting seat through bearings, one end of the X-axis adjustment screw is connected to the output shaft of the X-axis adjustment motor through a coupling, the X-axis guide rod is fixedly connected to the fixed mounting seat, and the two ends of the gantry are respectively provided with a first connecting member and a first sliding member, the first connecting member is screwed on the X-axis adjustment screw, and the first sliding member is screwed on the X-axis adjustment screw. The sliding sleeve is arranged on the X-axis guide rod, and the X-axis adjustment motor rotates to drive the gantry to move along the X-axis. The Y-axis adjustment mechanism is arranged on the bottom surface of the gantry. The Y-axis adjustment mechanism includes a Y-axis adjustment screw, a Y-axis guide rod and a bottom plate. The two ends of the Y-axis adjustment screw are respectively connected to two fixed seats through bearings. The two ends of the Y-axis guide rod are fixedly connected to the two fixed seats. One end of the Y-axis adjustment screw is connected to a driven gear, and the driven gear is meshed with a second driving gear. The second driving gear is connected to the output shaft of the Y-axis adjustment motor. The bottom plate is provided with a second connecting member, and the second connecting member is provided with a threaded hole that matches the Y-axis adjustment screw and a guide hole that matches the Y-axis guide rod. The second connecting member is screwed on the screw through the threaded hole and is slidably sleeved on the guide rod through the guide hole. The second camera is fixedly connected to the lower end surface of the bottom plate.

[0036] An unmanned substation status monitoring system based on image recognition comprises an image acquisition module, an image processing module, an alarm module and a controller, wherein four first cameras and a second camera are respectively connected to the image processing module via the image acquisition modules, the image processing module is connected to the controller, and the X-axis adjustment motor and the Y-axis adjustment motor are respectively connected to the controller.

[0037] The monitoring method of the monitoring system specifically comprises the following steps:

[0038] Step 1: Adjust the positions of the four first monitoring units at the four corners respectively to ensure that all the entrances and exits of the substation area are within the range of panoramic monitoring;

[0039] Step 2: The four first cameras collect the panoramic surveillance video of the scene and transmit it to the image acquisition module. The image acquisition module encodes and packages the panoramic surveillance video into an RTSP video stream and transmits it to the image processing module for image processing and storage.

[0040] Step 3: The image processing module decodes and analyzes the video stream to determine whether the person is not wearing a safety helmet, work clothes, or insulating gloves.

[0041] Step 4: When it is determined that the person is not wearing a safety helmet, work clothes, or insulating gloves, the video clip is saved, the video time period is recorded, and an alarm is sounded through the alarm module.

[0042] Step 5: The image processing module decodes and analyzes the video stream, determines the movement direction of the personnel, and transmits the entrance number of the personnel entering the substation and the personnel movement information to the controller. The controller first adjusts the second adjustable mechanism to move the second camera to the corresponding substation entrance, and adjusts the second adjustable mechanism in real time according to the personnel movement information.

[0043] Step 6: The second camera collects the tracking monitoring video and transmits it to the image acquisition module. The image acquisition module encodes and packages the tracking monitoring video into an RTSP video stream and transmits it to the image processing module for image processing.

[0044] Step 7, the image processing module decodes and analyzes the video stream to determine whether the person is not wearing a safety helmet, work clothes, or insulating gloves, and determines the direction of movement of the person, and then transmits the person's movement information to the controller, and the controller adjusts the second adjustable mechanism in real time to track the person.

[0045] In one embodiment, in the above steps, the image processing module determines and analyzes whether a person is not wearing a safety helmet, work clothes, or insulating gloves, and specifically includes the following steps:

[0046] Step a1, splicing and fusing the monitoring videos of the four first monitoring units A1-A4;

[0047] Step a2, performing image noise reduction preprocessing on the panoramic stitching fusion image;

[0048] Step a3: Segment the objects in the scene based on the mean shift algorithm;

[0049] Step a4, extracting the moving target using the background difference method;

[0050] Step a5: Detect the moving target using Hough transform to obtain a circular head area, a large rectangular torso area, and two small rectangular hand areas;

[0051] Step a6: extract the mean color of the helmet, the mean color of the work clothes, and the mean color of the insulating gloves at the corresponding position according to the installation angle and installation distance of the current camera;

[0052] Step a7, judge the mean color of the hard hat extracted in the previous step and the color matching degree of the hard hat, if they are less than the preset threshold, judge that the hard hat is not worn, judge the mean color of the work clothes extracted in the previous step and the color matching degree of the work clothes, if they are less than the preset threshold, judge that the work clothes are not worn, judge the mean color of the insulating gloves extracted in the previous step and the color matching degree of the insulating gloves, if they are less than the preset threshold, judge that the insulating gloves are not worn.

[0053] In one embodiment, the above-mentioned judgment and analysis of whether a person is not wearing a safety helmet, work clothes, or insulating gloves is first performed using a panoramic surveillance video, and when the second camera is adjusted into place, it is performed using a tracking surveillance video.

[0054] Step a1 specifically includes: collecting video streams of the four first monitoring units A1-A4 at the same time, decoding the video streams to obtain four RGB three-channel images, where the RGB three-channel images include an R channel image, a G channel image, and a B channel image respectively;

[0055] The RGB three-channel images of A1 and A3 are fused and stitched to form a first fused stitched image, the RGB three-channel images of A2 and A4 are fused and stitched to form a second fused stitched image, and the first fused stitched image and the second fused stitched image are fused and stitched to obtain a panoramic fused stitched image;

[0056] The fusion and stitching process of two images is as follows: find the overlapping area and the respective non-overlapping area between the two RGB three-channel images, use the feature matching method to obtain the feature relationship between the two RGB three-channel images, use the dynamic programming method to find the most suitable stitching line according to the similarity criterion, that is, the overlapping area between the two RGB three-channel images, remove the overlapping area, that is, the non-overlapping area, of each RGB three-channel image, construct the Alpha fusion image of the RGB three-channel image, the transparency value range of the Alpha fusion image is between 0 and 255, set the transparency value of the Alpha fusion image of the non-overlapping area of ​​the two RGB three-channel images to 255, set the transparency value of the Alpha fusion image of the overlapping area, and adopt a 255-0 gradient from inside to outside; use the OpenCV image processing method to divide the RGB three-channel image into R channel image, G channel image and B channel image, and then fuse the R channel image, G channel image and B channel image obtained after segmentation with the Alpha fusion image, and then splice them to obtain a fused splicing image, and keep the sum of the transparency values ​​of the overlapping area images to 255 during splicing;

[0057] Deghosting is performed on the overlapping areas of the fused mosaic.

[0058] In one embodiment, the specific method of de-ghosting is: calculate the homography matrix between the two images, the pixels in the two images correspond to the homography matrix, map the first image to the fused mosaic according to the correspondence between its pixels and the homography matrix, and then map the pixels of the second image to the pixels corresponding to the first image in the fused mosaic.

[0059] In one implementation, a drone is used to fly above the center of the substation monitoring area to collect a bird's-eye view image, and the bird's-eye view image is verified with the fused stitching image. The specific method is as follows:

[0060] S1. Using a point mapping method to perform distortion correction on the overhead image and the fused stitched image, and obtain the corrected distortion-free overhead image and the fused stitched image;

[0061] S2, performing perspective transformation on the undistorted top-view image to obtain a top-view image facing downward;

[0062] S3, extracting the first feature point in the distortion-free fused stitched image and the second feature point in the downward-looking overhead image, and obtaining the affine transformation matrix from each fused stitched image to the overhead image by using a straight-line linear transformation method;

[0063] S4. Performing an affine transformation on the fused and stitched images based on the affine transformation matrix of each fused and stitched image, and verifying the fused and stitched images after the affine transformation with the downward-looking overhead image.

[0064] In one embodiment, in the above steps, the image processing module determines and analyzes the moving direction of the person, specifically including the following steps:

[0065] Step b1, splicing and fusing the monitoring videos of the four first monitoring units A1-A4;

[0066] Step b2, performing image noise reduction preprocessing on the panoramic stitching fusion image;

[0067] Step b3: Segment the objects in the scene based on the mean shift algorithm;

[0068] Step b4, extracting the moving target using the background difference method;

[0069] Step b5, preferentially use the panoramic surveillance video of the first camera to analyze and determine the movement direction and position of the person, compare the current video frame with the video frame of the person at the entrance and exit positions, and obtain the movement information of the person. When the person leaves the current monitoring screen of the first camera, obtain the current time t1, and preferentially call up the panoramic surveillance videos of other first cameras to continue to analyze and determine the movement direction and position of the person, compare the current video frame with the video frame of the person at time t1, and obtain the movement information of the person. When all the monitoring screens of the first cameras cannot capture the person, obtain the current time t2, call up the tracking surveillance video of the second camera to continue to analyze and determine the movement direction and position of the person, compare the current video frame with the video frame of the previous second, and obtain the movement information of the person.

[0070] The above description is only a preferred embodiment of the present invention and does not limit the present invention in other forms. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An unmanned substation status monitoring device based on image recognition, characterized in that: It includes a panoramic monitoring mechanism and a tracking monitoring mechanism. The panoramic monitoring mechanism includes four first monitoring units, which are respectively arranged at the four corners of the unmanned substation. The tracking monitoring mechanism includes a second camera and a second adjustable mechanism. The second camera is arranged above the entire monitoring area through the second adjustable mechanism. The second adjustable mechanism includes an X-axis adjustment mechanism and a Y-axis adjustment mechanism.

2. The unmanned substation status monitoring device based on image recognition according to claim 1, characterized in that: The first monitoring unit includes a first column and a first camera. The first camera is installed on the column through a first adjustable mechanism. The first adjustable mechanism includes a ring gear, a ring track and a first mounting seat. The ring gear and the ring track are respectively connected to the column through a plurality of connecting rods, and the ring gear and the ring track are coaxially arranged. The first mounting seat includes a track connector, a first mounting plate and a first driving mechanism. The first driving mechanism includes a first driving motor and a first driving gear. The first driving motor is installed on the first mounting plate through a frame. The first driving gear is coaxially connected to the output shaft of the first driving motor. The first driving gear is meshed with the ring gear. The track connector is fixedly connected to the first mounting plate through a connecting rod. An arc-shaped slot is provided in the middle of the track connector. The track connector is slidably sleeved on the ring track through the arc-shaped slot. The first camera is connected to the lower end surface of the track connector.

3. The unmanned substation status monitoring device based on image recognition according to claim 1, characterized in that: The X-axis adjustment mechanism includes an X-axis adjustment screw, an X-axis guide rod and a gantry, wherein the X-axis adjustment screw and the X-axis guide rod are respectively arranged on the front and rear sides of the monitoring area through fixed mounting seats, and both ends of the X-axis adjustment screw are rotatably connected to the fixed mounting seat through bearings, and one end of the X-axis adjustment screw is connected to the output shaft of the X-axis adjustment motor through a coupling, and the X-axis guide rod is respectively fixedly connected to the fixed mounting seat, and the two ends of the gantry are respectively provided with a first connecting member and a first sliding member, the first connecting member is screwed on the X-axis adjustment screw, and the first sliding member is slidably sleeved on the X-axis guide rod, and the Y-axis adjustment mechanism is arranged on the bottom surface of the gantry, and the Y-axis adjustment mechanism includes Y The two ends of the Y-axis adjusting screw are rotatably connected to the two fixed seats through bearings respectively, and the two ends of the Y-axis guide rod are fixedly connected to the two fixed seats. One end of the Y-axis adjusting screw is connected to the driven gear, and the driven gear is meshed with the second driving gear. The second driving gear is connected to the output shaft of the Y-axis adjusting motor. A second connecting piece is provided on the base plate. The second connecting piece is provided with a threaded hole that cooperates with the Y-axis adjusting screw and a guide hole that cooperates with the Y-axis guide rod. The second connecting piece is screwed on the screw through the threaded hole and is slidably mounted on the guide rod through the guide hole. The second camera is fixedly connected to the lower end surface of the base plate.

4. An unmanned substation status monitoring system based on image recognition, characterized in that: The monitoring device comprises the monitoring device according to any one of claims 1 to 3, as well as an image acquisition module, an image processing module, an alarm module, and a controller, wherein the four first monitoring units and the second camera are connected to the image processing module via the image acquisition modules respectively, the image processing module is connected to the controller, and the X-axis adjustment mechanism and the Y-axis adjustment mechanism are connected to the controller respectively.

5. An unmanned substation status monitoring method based on image recognition, characterized in that: The specific steps include: Step 1: Adjust the positions of the four first monitoring units at the four corners respectively to ensure that all the entrances and exits of the substation area are within the range of panoramic monitoring; Step 2: The four first monitoring units collect panoramic monitoring videos of the scene and transmit them to the image acquisition module. The image acquisition module encodes and packages the panoramic monitoring videos into RTSP video streams and transmits them to the image processing module for image processing and storage. Step 3: The image processing module decodes and analyzes the video stream to determine whether the person is not wearing a safety helmet, work clothes, or insulating gloves. Step 4: When it is determined that the person is not wearing a safety helmet, work clothes, or insulating gloves, the video clip is saved, the video time period is recorded, and an alarm is sounded through the alarm module. Step 5: The image processing module decodes and analyzes the video stream, determines the movement direction of the personnel, and transmits the entrance number of the personnel entering the substation and the personnel movement information to the controller. The controller first adjusts the second adjustable mechanism to move the second camera to the corresponding substation entrance, and adjusts the second adjustable mechanism in real time according to the personnel movement information. Step 6: The second camera collects the tracking monitoring video and transmits it to the image acquisition module. The image acquisition module encodes and packages the tracking monitoring video into an RTSP video stream and transmits it to the image processing module for image processing. Step 7, the image processing module decodes and analyzes the video stream to determine whether the person is not wearing a safety helmet, work clothes, or insulating gloves, and determines the direction of movement of the person, and then transmits the person's movement information to the controller, and the controller adjusts the second adjustable mechanism in real time to track the person.

6. The unmanned substation status monitoring method based on image recognition according to claim 5, characterized in that: The image processing module determines and analyzes the movement direction of the person, which specifically includes the following steps: Step b1, decoding and segmenting the video into multiple video frames, and sorting the video frames by time to obtain orderly arranged video frames, Step b2, performing image noise reduction preprocessing on each video frame; Step b3: Segment the objects in the scene based on the mean shift algorithm; Step b4, extracting the moving target using the background difference method; Step b5, preferentially use the panoramic surveillance video of the first camera to analyze and determine the movement direction and position of the person, compare the current video frame with the video frame of the person at the entrance and exit positions, and obtain the movement information of the person. When the person leaves the current monitoring screen of the first camera, obtain the current time t1, and preferentially call up the panoramic surveillance videos of other first cameras to continue to analyze and determine the movement direction and position of the person, compare the current video frame with the video frame of the person at time t1, and obtain the movement information of the person. When all the monitoring screens of the first cameras cannot capture the person, obtain the current time t2, call up the tracking surveillance video of the second camera to continue to analyze and determine the movement direction and position of the person, compare the current video frame with the video frame of the previous second, and obtain the movement information of the person.

7. The unmanned substation status monitoring method based on image recognition according to claim 5, characterized in that: In the above steps, the image processing module determines and analyzes whether the person is not wearing a safety helmet, work clothes, or insulating gloves, and specifically includes the following steps: Step a1, splicing and fusing the monitoring videos of the four first monitoring units A1-A4; Step a2, performing image noise reduction preprocessing on the panoramic stitching fusion image; Step a3: Segment the objects in the scene based on the mean shift algorithm; Step a4, extracting the moving target using the background difference method; Step a5: Detect the moving target using Hough transform to obtain a circular head area, a large rectangular torso area, and two small rectangular hand areas; Step a6: extract the mean color of the helmet, the mean color of the work clothes, and the mean color of the insulating gloves at the corresponding position according to the installation angle and installation distance of the current camera; Step a7, judge the mean color of the hard hat extracted in the previous step and the color matching degree of the hard hat, if they are less than the preset threshold, judge that the hard hat is not worn, judge the mean color of the work clothes extracted in the previous step and the color matching degree of the work clothes, if they are less than the preset threshold, judge that the work clothes are not worn, judge the mean color of the insulating gloves extracted in the previous step and the color matching degree of the insulating gloves, if they are less than the preset threshold, judge that the insulating gloves are not worn.

8. The unmanned substation status monitoring method based on image recognition according to claim 7, characterized in that: Step a1 specifically includes: collecting video streams of the four first monitoring units A1-A4 at the same time, decoding the video streams to obtain four RGB three-channel images, where the RGB three-channel images include an R channel image, a G channel image, and a B channel image respectively; The RGB three-channel images of A1 and A3 are fused and stitched to form a first fused stitched image, the RGB three-channel images of A2 and A4 are fused and stitched to form a second fused stitched image, and the first fused stitched image and the second fused stitched image are fused and stitched to obtain a panoramic fused stitched image; The fusion and stitching process of two images is as follows: find the overlapping area and the respective non-overlapping area between the two RGB three-channel images, use the feature matching method to obtain the feature relationship between the two RGB three-channel images, use the dynamic programming method to find the most suitable stitching line according to the similarity criterion, that is, the overlapping area between the two RGB three-channel images, remove the overlapping area, that is, the non-overlapping area, of each RGB three-channel image, construct the Alpha fusion image of the RGB three-channel image, the transparency value range of the Alpha fusion image is between 0 and 255, set the transparency value of the Alpha fusion image of the non-overlapping area of ​​the two RGB three-channel images to 255, set the transparency value of the Alpha fusion image of the overlapping area, and adopt a 255-0 gradient from inside to outside; use the OpenCV image processing method to divide the RGB three-channel image into R channel image, G channel image and B channel image, and then fuse the R channel image, G channel image and B channel image obtained after segmentation with the Alpha fusion image, and then splice them to obtain a fused splicing image, and keep the sum of the transparency values ​​of the overlapping area images to 255 during splicing; Deghosting is performed on the overlapping areas of the fused mosaic.

9. The unmanned substation status monitoring method based on image recognition according to claim 8, characterized in that: Use a drone to fly above the center of the substation monitoring area to collect overhead images, and verify the overhead images with the fused stitching images. The specific method is as follows: S1. Using a point mapping method to perform distortion correction on the overhead image and the fused stitched image, and obtain the corrected distortion-free overhead image and the fused stitched image; S2, performing perspective transformation on the undistorted top-view image to obtain a top-view image facing downward; S3, extracting the first feature point in the distortion-free fused stitched image and the second feature point in the downward-looking overhead image, and obtaining the affine transformation matrix from each fused stitched image to the overhead image by using a straight-line linear transformation method; S4. Performing an affine transformation on the fused and stitched images based on the affine transformation matrix of each fused and stitched image, and verifying the fused and stitched images after the affine transformation with the downward-looking overhead image.

Citation Information

Patent Citations

  • Substation helmet wearing condition video monitoring system

    CN104902227A

  • Hierarchical joint-control intelligent monitoring method and system

    CN106791715A

  • Panorama image splicing method based on multichannel image fusion

    CN107203965A

  • Target object snapshot method and device and video monitoring equipment

    CN109922250A

  • Camera tracking control method and device

    CN111340857A

Cited By

  • Panoramic intelligent deployment and control sensing device in power grid system

    CN120512610A