Method and device for judging shielding of substation equipment
By using the difference calculation of RGB images and depth images in substation equipment, the problem of inaccurate occlusion recognition in low-light and complex environments is solved, accurate identification and early warning are achieved, and processing efficiency is improved.
Patent Information
- Application Number
- CN202310807086.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-30
- Publication Date
- 2024-12-31
AI Technical Summary
The prior art is difficult to accurately identify the occlusion situation of substation equipment under low light conditions and in complex environments, resulting in inaccurate identification.
By obtaining the RGB image and depth image of the target device without being blocked, the RGB value and depth value of each area are counted, and the difference is calculated from the real-time image to determine whether it is blocked, and the grid division method and change rate are used to use Calculate to improve recognition accuracy.
It realizes accurate identification of the occlusion situation of substation equipment in complex environments, timely issuance of occlusion warnings, obtains the position information of the occlusion, facilitates effective processing, reduces the influence of environmental factors such as light, and improves identification accuracy.
Smart Images
Figure CN119228718A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of substation equipment shielding detection, and in particular relates to a method and a device for determining substation equipment shielding. Background Art
[0002] Most existing substations use cameras to obtain target images and then obtain the target on-site working conditions. However, the camera position is fixed and the field of view is easily disturbed by unknown objects, such as Figure 1 As shown, identifying the obstruction and determining its location is key. On-site personnel can handle it in a timely manner based on the identification results to avoid dangerous situations.
[0003] Occlusion detection during camera shooting is an important problem in the field of computer vision. Its main purpose is to identify the occluded areas in the image and perform corresponding processing. Currently, commonly used occlusion detection technologies include methods based on deep learning, methods based on traditional computer vision technology, and methods based on optical flow.
[0004] The occlusion detection method based on deep learning is one of the most widely used methods. Its main idea is to achieve target detection by extracting and classifying features of images through convolutional neural networks (CNN). This method requires a large amount of training data sets and computing resources, and is not effective when the occlusion situation is complex.
[0005] The occlusion detection methods based on traditional computer vision technology mainly include image segmentation and background modeling. This method requires professional technicians to operate, and the effect is not good when the occlusion situation is complex.
[0006] The occlusion detection method based on optical flow mainly detects the motion trajectory of the object by calculating the pixel displacement between adjacent frames, thereby determining whether the object is occluded. This method does not work well under low light conditions and is easily affected by moving objects. Summary of the invention
[0007] The object of the present invention is to provide a method and device for determining shielding of substation equipment, so as to solve the problem of inaccurate identification of shielding conditions under low light conditions and in complex environments.
[0008] In order to solve the above technical problems, the present invention provides a method for determining whether substation equipment is blocked, comprising the following steps:
[0009] 1) Obtain the RGB image and depth image of the target device without occlusion, and call them the reference RGB image and reference depth image respectively, and divide the two images into regions in the same way; count the R value N in each region of the reference RGB image Rx 、G value NGx and B value N Bx ; Count the depth values of each area of the depth image Among them, x represents different areas;
[0010] 2) Collect the real-time RGB image and depth image of the target device, and call them the real-time RGB image and real-time depth image. The two images are processed in the same way as step 1) to obtain the R value N in the corresponding area of the real-time RGB image and the reference RGB image. p1Rx 、G value N p1Gx and B value N p1Bx ; and the depth values in the corresponding areas of the real-time depth image and the reference depth image Calculate N for each region of the image Rx and N p1Rx The difference between ΔR and N Gx and N p1Gx The difference between ΔG and N Bx and N p1Bx The difference ΔB and and The difference value Whether each divided area is blocked is determined according to the four difference values, thereby determining whether the target device is blocked.
[0011] The beneficial effects are as follows: in order to solve the problem of inaccurate recognition of the obstruction of the target device in a complex environment, the present invention uses the RGB image and depth image of the target device collected when it is not obstructed as the reference image, and performs the same area division as the real-time collected image, and counts the three RGB values of the RGB images of the real-time image and the reference image in each area, and the depth value of the depth image of the real-time image and the reference image in each area, and judges whether the target device is obstructed, and the position and obstruction method of the obstruction according to the difference of the above values in the corresponding area of the real-time image and the reference image. The present invention can issue an obstruction warning in time, which is convenient for on-site processing at the first time; it is conducive to obtaining the location information of the obstruction, and it is conducive to making a more effective processing method according to the actual situation; it can also change the reference image in real time to reduce the influence of environmental factors such as light, and improve the recognition accuracy.
[0012] Furthermore, the specific means of judging whether the divided area is blocked in step 2) is: if the difference values ΔR, ΔG, ΔB and If both are smaller than the set threshold, it is determined that the area in the real-time image is not blocked; otherwise, it is determined that the area is blocked.
[0013] The beneficial effect is that whether the target device in the whole image is blocked is judged according to whether the blocked areas occur in each divided area, ranging from small area to large area, which is conducive to the accurate identification of the blocked objects.
[0014] Furthermore, if ΔR, ΔG, ΔB and If ΔR, ΔG and ΔB of a region are all greater than or equal to the threshold, and If ΔR, ΔG and ΔB of a region are all smaller than the threshold, and If it is greater than or equal to the threshold, it is determined that the cause of occlusion in the area is occlusion by a similar object.
[0015] The beneficial effects are: according to whether occlusion occurs in each divided area, the image area occupied by the occlusion is found, which is conducive to the accurate identification of the occlusion and the acquisition of the specific location information of the occlusion, so as to facilitate more effective processing according to the actual situation.
[0016] Furthermore, the difference value of a certain parameter corresponding to a certain area in the real-time image and the comparison image refers to the change rate between the specific values of the parameter in the real-time image and the comparison image in the area.
[0017] Furthermore, when no occlusion occurs, the real-time image is used as a new comparison image.
[0018] The beneficial effect is that the difference between the divided areas is found by calculating the change rate, so that the result of judging whether there is an obstruction in each area is more accurate and efficient.
[0019] Furthermore, the R value of each region is obtained by averaging the R values of all pixels in the region; the G value of each region is obtained by averaging the G values of all pixels in the region; and the B value of each region is obtained by averaging the B values of all pixels in the region.
[0020] Furthermore, the depth value of each region is obtained by averaging the depth values of all pixels in the region.
[0021] Furthermore, the image is divided into regions by using a grid division method.
[0022] The beneficial effects are: making the size of the divided areas more even and regular, facilitating the calculation of the number of RGB image pixels and depth values in each area, and facilitating accurate finding of the differences in each divided area.
[0023] To solve the above technical problems, the present invention also provides a device for determining substation equipment obstruction, including a memory and a processor, wherein the processor is used to execute computer program instructions stored in the memory to implement the above-mentioned method for determining substation equipment obstruction.
[0024] The beneficial effect is that the device ensures the effective and reliable execution of a method for determining whether substation equipment is shielded. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 It is a schematic diagram of a camera acquiring a real-time picture of a substation in the prior art of the present invention;
[0026] Figure 2 It is a real-time comparison chart of the target object of the present invention;
[0027] Figure 3 is a flow chart of determining whether a target object is blocked according to the present invention;
[0028] Figure 4 It is a flow chart of obstacle blocking position determination of the present invention. DETAILED DESCRIPTION
[0029] The basic concept of the present invention is: the RGB image and depth image of the target device that are not obstructed captured by the camera are used as the reference images and the same area division is performed respectively, and the average values of the three RGB values in each area of the RGB image and the average values of all depth values in each area of the depth image are respectively counted; the two cameras capture the real-time image of the target device again, and perform the same processing as above, and judge whether the target device is obstructed, as well as the position and obstruction mode of the obstruction according to the difference between the above values in the corresponding area of the real-time image and the reference image. Based on this concept, a method for determining the obstruction of substation equipment and a device for determining the obstruction of substation equipment of the present invention can be realized.
[0030] The present invention will be described in detail below with reference to the accompanying drawings and method embodiments.
[0031] Method Example:
[0032] A method for determining shielding of substation equipment according to the present invention is shown in the flowchart as follows: Figure 3 As shown, the specific implementation steps include:
[0033] Step 1: Both the depth camera and the color camera are connected to the image processing component, and the depth data and color image are synchronously collected and transmitted to the image processing component; in the substation, the color camera and the depth camera respectively obtain an RGB image and a depth image of the target object that are not blocked as the reference image, and the RGB image and the depth image are divided into the same grid (region division), and the average value N of the three RGB values in each grid of the RGB image is respectively counted. Rx , N Gx and N Bx , calculate the average of all depth values contained in each grid of the depth image
[0034] Wherein, the value of x is the number of grids; in this embodiment, a 5*5 grid is used for grid division, so 25 groups of statistical values are used as a control group, such as Figure 2 shown.
[0035] Step 2: record the depth image and color image obtained by the two cameras in real time as p1 and p2; according to the principle of step 1, divide the depth image and color image obtained in real time in the same grid division method to obtain 25 sets of real-time data: the average value N of the three RGB values in each grid of the real-time RGB image p1Rx 、N p1Gx and N p1Bx ; and the average value of all depth values contained in each grid of the real-time depth image Calculate the rate of change of these values respectively, and set ΔR, ΔG, ΔB and The calculation formula of the change rate is as follows:
[0036]
[0037]
[0038]
[0039]
[0040] Among them, ΔR, ΔG and ΔB are the rates of change of the average values of the three values of RGB respectively; is the rate of change of the average value of the depth value; in this embodiment, a 5% rate of change is set as the threshold value based on the on-site conditions and a large number of experiments, and an image is acquired every 1 minute for detection.
[0041] Step 3: If ΔR, ΔG, ΔB and are greater than or equal to the threshold, it is determined that occlusion occurs, and according to The relationship between the size of the threshold and the cause of the occlusion is determined to be near occlusion; if ΔR, ΔG and ΔB are greater than or equal to the threshold, and If it is less than the threshold, it is determined that occlusion occurs, and according to The relationship between the size of the threshold and the cause of the occlusion is judged to be distant occlusion; if ΔR, ΔG and ΔB are less than the threshold, If it is greater than or equal to the threshold, it is determined that occlusion occurs, and according to The relationship between the size of the threshold and the cause of the occlusion is determined to be occlusion by a similar object; if ΔR, ΔG, ΔB and If all are less than the threshold, it is determined that no occlusion occurs. Because the environment may change slightly due to weather and other reasons, the real-time image is used as a new reference image, and step 1 is repeated to obtain 25 new reference maps for subsequent judgment of whether occlusion occurs. The flow chart is as follows: Figure 4 shown.
[0042] Among them, according to ΔR, ΔG, ΔB and The position with the largest rate of change can be used to determine the approximate direction and distance of the obstruction and provide early warning information.
[0043] Step 4: After the image processing component detects the obstruction, it sends the warning signal to the local warning device, and at the same time sends the warning signal to the network alarm device through the network for remote alarm.
[0044] This embodiment divides the image by grid division, so that each divided area has the same size. As other implementation methods, other methods that are conducive to identifying regional differences can be selected to divide the image, and the size of each area can be inconsistent.
[0045] This embodiment uses the method of calculating the rate of change of each average value of the divided area to characterize the difference between a grid in the real-time image and the corresponding grid in the reference image. As other implementation methods, other methods can also be selected to characterize the difference between the two, such as calculating the difference in parameters within corresponding grids between the two images for judgment.
[0046] In summary, the present invention can remotely and in real time obtain the occlusion situation of the target object image on site, automatically perform occlusion judgment, and issue occlusion warning in time to facilitate on-site processing at the first time; it can also judge whether it is lens occlusion, occlusion on the target object, or occlusion by an unknown object in the air, so as to obtain the position information of the occlusion, which is conducive to making more effective processing methods according to actual conditions; it can also change the reference image in real time to reduce the influence of environmental factors such as light to a certain extent.
[0047] Device Example:
[0048] A device for determining whether a substation device is blocked by electricity includes a memory and a processor, wherein the processor is used to execute computer program instructions stored in the memory to implement a method for determining whether a substation device is blocked by electricity as described in the method embodiment of the present invention. The processor may be a processing device such as a programmable logic device FPGA. The memory may be a variety of memories that use electrical energy to store information, such as RAM, ROM, etc., or other types of memories.
Claims
1. A method for determining shielding of substation equipment, characterized in that: The steps include: 1) Obtain the RGB image and depth image of the target device without occlusion, and call them the reference RGB image and reference depth image respectively, and divide the two images into regions in the same way; count the R value N in each region of the reference RGB image Rx 、G value N Gx and B value N Bx ; Count the depth values of each area of the depth image Among them, x represents different areas; 2) Collect the real-time RGB image and depth image of the target device, and call them the real-time RGB image and real-time depth image. The two images are processed in the same way as step 1) to obtain the R value N in the corresponding area of the real-time RGB image and the reference RGB image. p1Rx 、G value N p1Gx and B value N p1Bx ; and the depth values in the corresponding areas of the real-time depth image and the reference depth image Calculate N for each region of the image Rx and N p1Rx The difference between ΔR and N Gx and N p1Gx The difference between ΔG and N Bx and N p1Bx The difference ΔB and and The difference value Whether each divided area is blocked is determined according to the four difference values, thereby determining whether the target device is blocked.
2. The method for determining whether substation equipment is blocked according to claim 1, characterized in that: The specific means of judging whether the divided area is blocked in step 2) is: if the difference values ΔR, ΔG, ΔB and If both are smaller than the set threshold, it is determined that the area in the real-time image is not blocked; otherwise, it is determined that the area is blocked.
3. The method for determining whether substation equipment is blocked according to claim 2, characterized in that: If the ΔR, ΔG, ΔB and If ΔR, ΔG and ΔB of a region are all greater than or equal to the threshold, and If ΔR, ΔG and ΔB of a region are all smaller than the threshold, and If it is greater than or equal to the threshold, it is determined that the cause of occlusion in the area is occlusion by a similar object.
4. The method for determining whether substation equipment is blocked according to claim 1 or 2, characterized in that: The difference value of a certain parameter corresponding to a certain area in the real-time image and the comparison image refers to the change rate between the specific values of the parameter in the real-time image and the comparison image in the area.
5. The method for determining whether substation equipment is blocked according to claim 3, characterized in that: When no occlusion occurs, the real-time image is used as a new comparison image.
6. The method for determining whether substation equipment is blocked according to claim 1, characterized in that: The R value of each region is obtained by averaging the R values of all pixels in the region; the G value of each region is obtained by averaging the G values of all pixels in the region; and the B value of each region is obtained by averaging the B values of all pixels in the region.
7. The method for determining whether substation equipment is shielded according to claim 1, characterized in that: The depth value of each area is obtained by averaging the depth values of all pixels in the area.
8. The method for determining whether substation equipment is shielded according to claim 1, characterized in that: The image is divided into regions using grid division.
9. A device for determining whether substation equipment is blocked, characterized in that: It comprises a memory and a processor, wherein the processor is used to execute computer program instructions stored in the memory to implement the method for determining substation equipment shielding as described in any one of claims 1 to 8.