Transformer substation inspection method and system based on image recognition

Abstract

The invention discloses a transformer substation inspection method and system based on image recognition, and relates to the technical field of image processing, and the method comprises the steps: dividing a regional image of a transformer substation into a plurality of image blocks, and determining the definition of each image block; determining a clear boundary according to the definition; determining a to-be-collected area according to the clear boundary; acquiring a wireless sensing range of the state acquisition sensor; planning a moving path of the mobile acquisition device; controlling the mobile acquisition device to move according to the moving path; and identifying the operation state of the transformer substation according to the clear image block, the image of the to-be-collected area and the transformer substation state data. According to the method, the image of the transformer substation is subjected to definition recognition, so that the fuzzy area under the influence of illumination is determined, then the mobile acquisition device is controlled to reacquire the image in the fuzzy area and the state data of the equipment in the transformer substation, and a relatively accurate recognition result is obtained based on the image and the state data.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a substation inspection method and system based on image recognition. Background Art

[0002] Substations are facilities that adjust the voltage during the transmission process. In order to adapt to long-distance transmission, the voltage of the power plant output is usually increased to tens or even hundreds of kilovolts. Before it is transmitted to power users or residents' homes, the voltage is reduced to hundreds of volts. Substations are required to participate in both voltage increase and decrease, so substations play a very important role in the safe and smooth transmission of power.

[0003] There are many kinds of equipment in the substation, including transformers, switches, etc. These equipment need to work continuously, so it is inevitable that failures will occur during operation. However, due to the strong electromagnetic field in the substation and the serious consequences of failure, the substation usually does not arrange personnel to be on duty on site, but obtains substation information in real time through remote monitoring. During the remote monitoring process, multiple cameras will be installed in the substation, and video monitoring will enable the on-duty personnel to understand the situation on site of the substation. The images obtained through video monitoring can also automatically identify the operating status of the equipment in the substation, replacing manual identification and improving the automation level of substation inspection.

[0004] However, due to the influence of lighting, images obtained through video surveillance may have blurred areas in some areas. Since effective features cannot be extracted from these blurred areas, the accuracy of the recognition results cannot be ensured, which seriously threatens the effectiveness of remote monitoring of substations. Summary of the invention

[0005] The embodiments of the present application provide a substation inspection method and system based on image recognition, which are used to solve the problem in the prior art of inaccurate equipment operation status recognition results caused by blurred images obtained by video surveillance.

[0006] On the one hand, an embodiment of the present application provides a substation inspection method based on image recognition, including: Get regional images of the substation; Divide the regional image into a plurality of image blocks, analyze each image block, and determine the clarity of each image block; Determining a clear boundary of the regional image according to the clarity of the image block, the clear boundary being a boundary between a clear image block and a blurred image block in the regional image; Determine the area to be collected in the substation based on clear boundaries; Obtain the wireless sensing range of each status acquisition sensor in the substation and the moving speed of the mobile acquisition device; Plan the moving path of the mobile acquisition device in the substation so that the mobile acquisition device can collect images of all areas to be collected during the movement, and can also collect substation status data collected by the status collection sensor within the wireless sensing range; Control the mobile acquisition device to move along the moving path to acquire images of the area to be acquired and the status data of the substation; The operating status of the substation is identified based on the clear image blocks in the regional image, the image of the area to be collected, and the substation status data.

[0007] On the other hand, the embodiment of the present application also provides a substation inspection system based on image recognition, including: An image acquisition module, used to acquire a regional image of the substation; A clarity analysis module is used to divide the regional image into a plurality of image blocks, analyze each image block, and determine the clarity of each image block; A boundary determination module, used to determine a clear boundary of the regional image according to the clarity of the image block, the clear boundary being a dividing line between a clear image block and a blurred image block in the regional image; The area determination module is used to determine the area to be collected in the substation according to clear boundaries; A range acquisition module is used to obtain the wireless sensing range of each status acquisition sensor in the substation and the moving speed of the mobile acquisition device; A path planning module is used to plan the moving path of the mobile acquisition device in the substation according to the moving speed, so that the mobile acquisition device can collect images of all areas to be collected during the movement, and can also collect substation status data collected by the status collection sensor within the wireless sensing range; A data acquisition module is used to control the mobile acquisition device to move along the moving path to acquire images of the area to be acquired and the status data of the substation; The state recognition module is used to identify the operating state of the substation based on the clear image blocks in the regional image, the image of the area to be collected, and the substation state data.

[0008] The substation inspection method and system based on image recognition in this application have the following advantages: By performing clarity recognition on the substation image, the blurred area under the influence of light can be determined, and then the mobile acquisition device can be controlled to re-capture images and status data of equipment in the substation in the blurred area. Based on the images of the clear area, the re-captured images and status data, a more accurate recognition result can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0010] Figure 1 A flowchart of a substation inspection method based on image recognition provided in an embodiment of the present application. DETAILED DESCRIPTION

[0011] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0012] Figure 1 A flowchart of a substation inspection method based on image recognition provided in an embodiment of the present application. An embodiment of the present application provides a substation inspection method based on image recognition, including: S100, obtaining a regional image of a substation.

[0013] For example, a substation usually has multiple fixed surveillance cameras installed, and each surveillance camera faces a specific direction of the substation. Due to the limitation of the field of view, the regional image captured by each surveillance camera can only cover part of the substation. Even if there is a certain overlap in the areas covered by multiple surveillance cameras, it is impossible to obtain effective information for accurately identifying the operating status of the equipment in the substation from the regional image alone. Therefore, in addition to obtaining the regional image through the surveillance camera, it is also necessary to deploy at least one mobile collection device in the substation. The mobile collection device can be a wheeled or tracked robot, or a multi-rotor drone. These mobile collection devices can collect images of areas in the substation corresponding to the blurred areas in the regional image, thereby providing effective data for accurate identification of the operating status of the equipment.

[0014] S110, dividing the regional image into a plurality of image blocks, analyzing each image block, and determining the clarity of each image block.

[0015] Exemplarily, the regional image can be divided into multiple image blocks in rows and / or columns, and the size of each image block should remain the same, and the image block should be small enough to make the determined clear boundary sufficiently accurate, but the image block should not be too small to avoid generating too much calculation. In the embodiment of the present application, when the monitoring camera uses a horizontal or smaller downward angle to collect and obtain the regional image, the area closer to the monitoring camera in the regional image has a higher definition, and conversely, the area farther from the monitoring camera has a lower definition. According to this linear change trend, there must be a vertical or inclined straight line in the regional image, and on both sides of the straight line are clear and blurred image blocks, respectively. Therefore, the image block in the present application is set as a vertical area running through the top and bottom of the regional image, and the width of the vertical area is a set number of pixels, such as 10 pixels or 20 pixels. Adjacent image blocks can have a certain overlap to completely divide the entire regional image into multiple image blocks.

[0016] In the embodiment of the present application, the Tenengrad gradient method, the Laplacian gradient method or the variance method can be used to determine the clarity of each image block. Among them, the Tenengrad gradient method and the Laplacian gradient method are two methods for calculating the gradient of an image in the horizontal and vertical directions. The higher the gradient value, the higher the clarity. The variance method is used to measure the degree of difference between pixels in an image block. The larger the variance, the higher the clarity.

[0017] S120, determining a clear boundary of the regional image according to the clarity of the image block, where the clear boundary is a boundary between a clear image block and a blurred image block in the regional image.

[0018] Exemplarily, after determining the clarity of each image block, the clarity of each image block is compared with a clarity threshold, and the first image block whose clarity is lower than the clarity threshold in a set direction is taken as a blurred image block, while the image block before the blurred image block in the set direction is taken as a clear image block.

[0019] The above-mentioned set direction is the direction away from the surveillance camera. Since the distance between the image block and the surveillance camera along this direction gradually increases, according to the perspective principle, the pixels representing the equipment in the substation in these image blocks will gradually decrease. When the number of pixels is reduced to a certain extent, the subsequent state recognition model cannot extract enough features, and thus cannot obtain accurate recognition results. Therefore, it is necessary to re-collect the image of the fuzzy area with the help of a mobile acquisition device. In the process of identifying the clear boundary, the present application will identify the first image block from one end of the regional image according to the set direction. If the clarity of the first image block is not less than the clarity threshold, then analyze the next image block until the clarity of a certain image block is lower than the clarity threshold. The image block is the first blurred image block after the clear boundary. Correspondingly, the last image block in front of the image block is a clear image block, and the boundary between the two image blocks is the clear boundary. If there is an overlap between two adjacent image blocks, the edge of the blurred image block close to the clear image block can be used as a clear boundary. After the first blurred image block is identified, the subsequent image blocks do not need to be further analyzed, and the final clear boundary is directly determined.

[0020] Since the clarity of regional images is greatly affected by lighting, regional images captured by the same surveillance camera under different lighting conditions also have different clarity. Therefore, the clear boundaries determined at different times are also different, which leads to differences in the moving paths of the mobile acquisition device.

[0021] S130, determining the area to be collected in the substation according to the clear boundary.

[0022] Exemplarily, each regional image corresponds to an area in the substation. After determining the clear boundaries of the regional image, the blurred image blocks in the set direction and the areas corresponding to the image blocks behind are taken as the areas to be collected. The areas to be collected corresponding to the image blocks in all regional images constitute the areas to be collected of the entire substation.

[0023] Specifically, although the regional image is a plane image, it reflects the spatial information. The clear boundary determined on the regional image represents the actual spatial area. By determining the position of the clear boundary on the regional image and mapping the position in the area corresponding to the regional image, the boundary between the clear area and the fuzzy area can be determined. The area behind the boundary in the set direction is the area to be collected. After the clear boundary is determined for each regional image, the area to be collected in the entire substation can be summarized.

[0024] S140, obtaining the wireless sensing range of each status collection sensor in the substation and the moving speed of the mobile collection device.

[0025] Exemplarily, in the embodiment of the present application, a plurality of state acquisition sensors are installed in each device of the substation. These state acquisition sensors can collect data such as the current, voltage, vibration and temperature of the equipment, and these data can be sent wirelessly. Although the range of the substation is not very large, due to the presence of a strong electromagnetic field, even if wireless transmission is adopted, the data collected by the state acquisition sensor cannot be accurately transmitted to the receiving device at a long distance. Therefore, the present application uses a mobile acquisition device to receive data. In order to improve the quality of wireless data communication, the present application uses Bluetooth technology to connect the state acquisition sensor and the mobile acquisition device, so as to transmit the collected data to the mobile acquisition device after the two establish a wireless connection.

[0026] Regardless of whether Bluetooth, WiFi or other wireless communication technologies are used, there is a corresponding communication distance. Before the method of the embodiment of the present application is executed, the substation can be three-dimensionally modeled. Specifically, the size of the substation site and the size and position data of each device can be collected to establish a three-dimensional model. In the three-dimensional model, the position of the monitoring camera and the position of the status acquisition sensor can be added. By analyzing the position, direction and field of view of the monitoring camera, each area to be collected can be determined in combination with the position of the clear boundary in the regional image. After drawing a sphere with the position of the status acquisition sensor as the center and the communication distance as the radius, the area enclosed by the sphere is the wireless sensing range. The mobile acquisition device needs to be within the wireless sensing range to obtain the data collected by the status acquisition sensor.

[0027] S150, planning a moving path of the mobile collection device in the substation, so that the mobile collection device can collect images of all areas to be collected during the movement, and can also collect substation status data collected by the status collection sensor within the wireless sensing range.

[0028] Exemplarily, when planning a moving path, multiple groups of collection points are determined in advance, and each group of collection points ensures that the mobile collection device can collect images of all areas to be collected during the movement, and can also collect substation status data collected by the status collection sensor within the wireless sensing range. The moving time of two adjacent collection points in each group is determined according to the moving speed, and multiple groups of collection points are screened with the goal of shortest total moving time to obtain a moving path.

[0029] When setting the collection points, a certain order is required. Specifically, a certain area to be collected can be selected as the starting point, and multiple areas to be collected can be connected in series in a certain direction, and then multiple areas to be collected can be connected. The collection points will be connected in sequence according to the order of the areas to be collected, so that when the mobile collection device moves according to the position of the collection point, unnecessary movement can be reduced as much as possible.

[0030] Furthermore, the total moving time is the sum of the movements in a set of collection points. An objective function is established with the shortest total moving time as the goal, and the moving path is obtained by solving the objective function. The objective function is subject to constraints in the solution process. The constraints include collection range constraints and perception range constraints. The collection range constraints indicate that the mobile collection device can collect images of all areas to be collected during the movement. The perception range constraints indicate that when the mobile collection device is located at the collection point, it can collect substation status data collected by the status collection sensor within the wireless perception range.

[0031] The above objective function can be expressed as:

[0032] in, f represents the objective function, min represents the minimum value, M Indicates the number of all collection points in the substation, i Indicates the labels of all the collection points. Indicates that from i The collection point moves to i+ The time required for one collection point is expressed as:

[0033] in, Indicates i The collection points and i+ The distance between 1 collection points, v Indicates the movement speed.

[0034] The collection range constraint is expressed as:

[0035] in, M n Indicates that the mobile collection device is collecting n The number of collection points to be set when there are areas to be collected. n =1,2,…, N , N Indicates the number of areas to be collected. M 1+ M 2+…+ M N = M , i n Indicates n The number of the collection points in the area to be collected, Indicates that the mobile collection device is n In the area to be collected i n The area width corresponding to the image collected when there are 1 collection point isw n Indicates n The width of the area to be collected.

[0036] The perception range constraint is expressed as:

[0037] in, Indicates that the mobile collection device is n In the area to be collected i n The collection points and s The distance between the state acquisition sensors, r s Indicates s The wireless sensing range of a state acquisition sensor.

[0038] Furthermore, a genetic algorithm can be used to solve the objective function. A genetic algorithm is a bionic algorithm that simulates the characteristics of biological intergenerational inheritance. Through the three steps of selection, crossover and mutation, it can generate as many collection points as possible that meet the constraints, thereby greatly improving the diversity of the collection points, and ensuring that the collection points finally screened are as optimal as possible.

[0039] S160, controlling the mobile acquisition device to move along the moving path to acquire images of the area to be acquired and the status data of the substation.

[0040] For example, since the mobile path is a plurality of collection points connected in sequence, the mobile collection device only needs to move to each collection point in sequence, and then collect images and substation status data at each collection point. After completing the collection work of the entire mobile path, the mobile collection device can return to the specified position and send the collected images and substation status data to the processing host by wireless communication.

[0041] S170, identifying the operating status of the substation according to the clear image blocks in the regional image, the image of the area to be collected, and the substation status data.

[0042] Exemplarily, a state recognition model is deployed in the processing host. When the mobile acquisition device sends clear image blocks in the regional image, images of the area to be acquired, and substation status data to the processing host, the state recognition model therein can recognize these images and data to determine the operating status of the substation.

[0043] The above-mentioned state recognition model can be established based on a neural network. Before use, the state recognition model needs to be trained, verified and tested using a training set, a validation set and a test set. After the recognition accuracy meets the requirements, the above-mentioned image and substation state data can be input into the state recognition model to obtain the recognition result of whether the equipment in the substation is normal.

[0044] The embodiment of the present application also provides a substation inspection system based on image recognition, the system comprising: An image acquisition module, used to acquire a regional image of the substation; A clarity analysis module is used to divide the regional image into a plurality of image blocks, analyze each image block, and determine the clarity of each image block; A boundary determination module, used to determine a clear boundary of the regional image according to the clarity of the image block, the clear boundary being a dividing line between a clear image block and a blurred image block in the regional image; The area determination module is used to determine the area to be collected in the substation according to clear boundaries; A range acquisition module is used to obtain the wireless sensing range of each status acquisition sensor in the substation and the moving speed of the mobile acquisition device; A path planning module is used to plan the moving path of the mobile acquisition device in the substation according to the moving speed, so that the mobile acquisition device can collect images of all areas to be collected during the movement, and can also collect substation status data collected by the status collection sensor within the wireless sensing range; A data acquisition module is used to control the mobile acquisition device to move along the moving path to acquire images of the area to be acquired and the status data of the substation; The state recognition module is used to identify the operating state of the substation based on the clear image blocks in the regional image, the image of the area to be collected, and the substation state data.

[0045] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0046] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A substation inspection method based on image recognition, characterized in that: include: Get regional images of the substation; Dividing the regional image into a plurality of image blocks, analyzing each of the image blocks, and determining the clarity of each of the image blocks; Determining a clear boundary of the regional image according to the clarity of the image block, wherein the clear boundary is a boundary between the clear image block and the blurred image block in the regional image; Determine the area to be collected in the substation according to the clear boundary; Obtain the wireless sensing range of each status acquisition sensor in the substation and the moving speed of the mobile acquisition device; Planning a moving path of the mobile acquisition device in the substation according to the moving speed, so that the mobile acquisition device can acquire images of all the areas to be acquired during the movement, and can also acquire the substation status data acquired by the status acquisition sensor within the wireless sensing range; Controlling the mobile acquisition device to move along the moving path to acquire the image of the area to be acquired and the substation status data; The operation status of the substation is identified according to the clear image blocks in the regional image, the image of the area to be collected and the substation status data.

2. The substation inspection method based on image recognition according to claim 1 is characterized in that: The clarity of each image block is determined by using a Tenengrad gradient method, a Laplacian gradient method or a variance method.

3. The substation inspection method based on image recognition according to claim 1 is characterized in that: After determining the clarity of each image block, the clarity of each image block is compared with a clarity threshold, and the first image block whose clarity is lower than the clarity threshold in a set direction is taken as the blurred image block, and the image block before the blurred image block in the set direction is taken as the clear image block.

4. The substation inspection method based on image recognition according to claim 3 is characterized in that: Each of the regional images corresponds to an area in the substation. After determining the clear boundary of the regional image, the image block blurred in the set direction and the area corresponding to the image block behind are taken as the area to be collected, and the areas to be collected corresponding to all the image blocks constitute the area to be collected of the substation.

5. The substation inspection method based on image recognition according to claim 1 is characterized in that: When planning the moving path, multiple groups of collection points are predetermined, and each group of the collection points ensures that the mobile collection device can collect images of all the areas to be collected during the movement, and can also collect the substation status data collected by the status collection sensor within the wireless sensing range. The moving time of two adjacent collection points in each group is determined according to the moving speed, and the multiple groups of collection points are screened with the goal of minimizing the total moving time to obtain the moving path.

6. The substation inspection method based on image recognition according to claim 5 is characterized in that: The total moving time is the sum of the movements in a group of the collection points. An objective function is established with the shortest total moving time as the goal, and the moving path is obtained by solving the objective function. The objective function is subject to constraints during the solution process. The constraints include collection range constraints and perception range constraints. The collection range constraints indicate that the mobile collection device can collect images of all the areas to be collected during the movement, and the perception range constraints indicate that when the mobile collection device is located at the collection point, it can collect the substation status data collected by the status collection sensor within the wireless perception range.

7. The substation inspection method based on image recognition according to claim 6 is characterized in that: A genetic algorithm is used to solve the objective function.

8. The substation inspection method based on image recognition according to claim 1 is characterized in that: The clear image blocks in the regional image, the image of the area to be collected and the substation status data are input into a pre-trained status recognition model to determine the operating status of the substation.

9. A system using the substation inspection method based on image recognition according to any one of claims 1 to 8, characterized in that: include: An image acquisition module, used to acquire a regional image of the substation; A clarity analysis module, used for dividing the regional image into a plurality of image blocks, analyzing each of the image blocks, and determining the clarity of each of the image blocks; A boundary determination module, used to determine a clear boundary of the regional image according to the clarity of the image block, wherein the clear boundary is a boundary between the clear image block and the blurred image block in the regional image; An area determination module, used to determine the area to be collected in the substation according to the clear boundary; A range acquisition module is used to obtain the wireless sensing range of each status acquisition sensor in the substation and the moving speed of the mobile acquisition device; A path planning module, used for planning the moving path of the mobile acquisition device in the substation according to the moving speed, so that the mobile acquisition device can collect images of all the areas to be collected during the movement, and can also collect the substation status data collected by the status acquisition sensor within the wireless sensing range; A data acquisition module, used for controlling the mobile acquisition device to move along the moving path to acquire the image of the area to be acquired and the substation status data; The state recognition module is used to recognize the operation state of the substation according to the clear image blocks in the regional image, the image of the area to be collected and the substation state data.

Images