Transformer substation inspection path optimization method based on binocular vision navigation

Abstract

The invention discloses a substation inspection path optimization method based on binocular vision navigation, and relates to the field of substation inspection. The problem that a traditional path planning method based on visual navigation consumes long time in the planning process is solved. Performing initialization preprocessing on the inspection map data; arranging nodes on the visual interaction planning inspection map; performing data acquisition on the intelligent substation by using a binocular vision system to obtain image data of equipment; the collected image data is preprocessed; extracting key feature points of the equipment from the image data by adopting a feature extraction algorithm; based on the binocular vision principle, three-dimensional reconstruction of the equipment is realized by calculating the parallax between the two cameras; according to the three-dimensional re-seeing model of the equipment, a routing inspection path is determined by adopting a path planning algorithm in combination with node layout, so that the path length is shortest and each part of the equipment is covered; the planned path is optimized, and path optimization is completed. The method is mainly used for transformer substation inspection path optimization.

Description

Technical Field

[0001] The invention relates to the field of substation inspection. Background Art

[0002] With the rapid development of the power industry, substations, as an important part of the power system, play a key role in power transmission, distribution, and conversion. However, traditional manual inspection methods have problems such as low efficiency, high workload, and high safety risks.

[0003] Visual navigation is a new navigation technology that uses sensor elements to perceive the surrounding environment and uses the collected data as the basis for path planning of inspection equipment. Visual navigation mainly includes image preprocessing, target extraction and tracking, information fusion and other application technologies. It does not require the setting of clear information query restrictions or the consistency of statistical data samples. Therefore, in actual application, visual navigation path planning technology has stronger adaptability.

[0004] In traditional methods, navigation path planning based on deep reinforcement learning uses a reward value model to analyze the integrity of the navigation path, and uses a neural network structure to complete the on-demand allocation of navigation nodes. The navigation ranging method based on binocular vision technology follows the principle of real-time ranging, defines the interval distance between adjacent navigation nodes, and compares the preset point objects with the actual inspection nodes to determine the degree of deviation between the two. However, the application capabilities of the above two methods are limited. In the process of implementing substation inspections, it is difficult for the actual inspection nodes of the robot to completely match the preset nodes. This requires the control host to consume a certain amount of time to identify the node objects, resulting in a long path planning process, which does not meet the needs of quickly extracting navigation paths. Summary of the invention

[0005] The purpose of the present invention is to solve the problem that the planning process of the traditional path planning method based on vision navigation is time-consuming. The present invention provides a substation inspection path optimization method based on binocular vision navigation.

[0006] The substation inspection path optimization method based on binocular vision navigation includes the following steps:

[0007] Initialization and preprocessing of inspection map data;

[0008] Visual interaction planning of node layout on inspection map;

[0009] Use binocular vision system to collect data from smart substations and obtain image data of equipment;

[0010] Preprocessing the collected image data;

[0011] Use feature extraction algorithm to extract key feature points of the device from image data;

[0012] Based on the principle of binocular vision, the three-dimensional reconstruction of the device is achieved by calculating the parallax between the two cameras;

[0013] According to the 3D re-model of the equipment, the path planning algorithm is used in combination with the node layout to determine the inspection path so that the path length is the shortest and covers every part of the equipment;

[0014] The planned path is optimized to complete the path optimization.

[0015] Preferably, the inspection map data initialization preprocessing is implemented as follows:

[0016] The inspection map data is described as a raster map and organized into a basic planning matrix.

[0017] Preferably, the implementation method of the node layout on the visual interactive planning inspection map is:

[0018] Set JPS pathfinding visual intersection planning nodes on the grid map, and use binocular vision technology to establish a JPS interactive recognition space in the specified inspection range.

[0019] Preferably, preprocessing the acquired image data includes image denoising and image enhancement.

[0020] Preferably, the key feature points of the device include edges and corners of the device.

[0021] Preferably, the DIJKSTRA algorithm is used to determine the shortest inspection path planning.

[0022] Advantages of the present invention:

[0023] Binocular vision technology has broad application prospects in intelligent substation inspection path planning. This paper studies the application of binocular vision technology in intelligent substation inspection path planning. By using binocular images obtained by binocular cameras, the depth information of objects in the scene can be calculated and converted into three-dimensional point cloud data. These data can help intelligent systems accurately detect and locate substation equipment, thereby providing a basis for inspection path planning.

[0024] The present invention establishes an accurate map model, which is beneficial for the staff to navigate according to the map model. At the same time, during the inspection of the smart substation, there are usually multiple target points that need to be inspected. In order to improve the inspection efficiency, these target points need to be optimized. The binocular vision system is used to detect the target points, and the target points are combined with the map model to achieve fast and accurate intelligent path planning, avoiding repeated walking caused by unreasonable inspection order. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a flow chart of the substation inspection path optimization method based on binocular vision navigation described in the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0028] The so-called binocular vision is a three-dimensional visual system designed to imitate the principle of human binocular vision. Through two cameras that are parallel to each other but have a certain distance between them, the binocular vision system can obtain the depth information of the object. In the substation inspection, the binocular vision system can be used to obtain the stereoscopic image of the substation equipment, and analyze the image to realize the automatic inspection path planning. The present invention provides the following implementation methods based on binocular vision, specifically:

[0029] Specific implementation method 1. Figure 1 The present embodiment describes a method for optimizing a substation inspection path based on binocular vision navigation, comprising the following steps:

[0030] Initialize preprocessing of inspection map data;

[0031] Visual interaction planning of node layout on inspection map;

[0032] Use binocular vision system to collect data from smart substations and obtain image data of equipment;

[0033] Preprocessing the collected image data;

[0034] Use feature extraction algorithm to extract key feature points of the device from image data;

[0035] Based on the principle of binocular vision, the three-dimensional reconstruction of the device is achieved by calculating the parallax between the two cameras;

[0036] According to the 3D re-model of the equipment, the path planning algorithm is used in combination with the node layout to determine the inspection path so that the path length is the shortest and covers every part of the equipment;

[0037] The planned path is optimized to complete the path optimization.

[0038] In this implementation, before path planning, it is necessary to initialize and preprocess the inspection map data, and the inspection range of the substation needs to be calibrated, so the inspection map data is initialized and preprocessed; the path planning should take into account factors such as the distance between devices, channel width, and device accessibility, so as to ensure the efficiency of the inspection process, and visually interact with the planning of node layout on the inspection map. Specifically, the DIJKSTRA algorithm is used to determine the shortest inspection path planning.

[0039] Binocular vision technology has broad application prospects in intelligent substation inspection path planning. This paper studies the application of binocular vision technology in intelligent substation inspection path planning. By using binocular images obtained by binocular cameras, the depth information of objects in the scene can be calculated and converted into three-dimensional point cloud data. These data can help the intelligent system accurately detect and locate substation equipment, thereby providing a data basis for inspection path planning and providing path optimization speed.

[0040] Specifically, when planning the inspection route, it is necessary to accurately map the substation, use lidar and other sensor technologies to obtain the three-dimensional point cloud data inside the substation, and then establish an accurate map model, which is conducive to the staff to navigate according to the map model. At the same time, during the inspection of smart substations, there are usually multiple target points that need to be inspected. In order to improve the inspection efficiency, these target points need to be optimized. Based on the Hopfield obstacle avoidance visual control technology, the binocular vision system is used to detect the target points, and the target points are combined with the map model to realize intelligent path planning and avoid repeated walking caused by unreasonable inspection sequence.

[0041] Furthermore, the inspection map data initialization preprocessing is implemented by describing the inspection map data as a grid map and arranging it into a basic planning matrix.

[0042] In the preferred implementation manner, the inspection map data is described as a grid map for data preprocessing to filter out valid selectable data, thus saving path planning time.

[0043] Furthermore, the implementation method of node layout on the visual interactive planning inspection map is as follows:

[0044] Set JPS pathfinding visual intersection planning nodes on the grid map, and use binocular vision technology to establish a JPS interactive recognition space in the specified inspection range.

[0045] In specific applications, for a specific node in the smart substation, the corresponding obstacle jump points are collected and the shortest distance between the interactive planning nodes is calculated.

[0046] Furthermore, the collected image data is preprocessed, including image denoising, image enhancement and other operations, to improve image quality.

[0047] Furthermore, the key feature points of the device include the edges and corners of the device. The device is accurately reconstructed in three dimensions through the above key feature points.

[0048] Although the present invention is described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. It should therefore be understood that many modifications may be made to the exemplary embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the various dependent claims and features described herein may be combined in a manner different from that described in the original claims. It should also be understood that features described in conjunction with individual embodiments may be used in other described embodiments.

Claims

1. A substation inspection path optimization method based on binocular vision navigation is characterized in that: The steps include: Initialization and preprocessing of inspection map data; Visual interaction planning of node layout on inspection map; Use binocular vision system to collect data from smart substations and obtain image data of equipment; Preprocessing the collected image data; Use feature extraction algorithm to extract key feature points of the device from image data; Based on the principle of binocular vision, the three-dimensional reconstruction of the device is achieved by calculating the parallax between the two cameras; According to the 3D re-model of the equipment, the path planning algorithm is used in combination with the node layout to determine the inspection path so that the path length is the shortest and covers every part of the equipment; The planned path is optimized to complete the path optimization.

2. The substation inspection path optimization method based on binocular vision navigation according to claim 1 is characterized in that: The implementation method of initialization preprocessing of inspection map data is as follows: The inspection map data is described as a raster map and organized into a basic planning matrix.

3. The substation inspection path optimization method based on binocular vision navigation according to claim 1 is characterized in that: The implementation method of node layout on the visual interactive planning inspection map is as follows: Set JPS pathfinding visual intersection planning nodes on the grid map, and use binocular vision technology to establish a JPS interactive recognition space in the specified inspection range.

4. The substation inspection path optimization method based on binocular vision navigation according to claim 1 is characterized in that: The preprocessing of the acquired image data includes image denoising and image enhancement.

5. The substation inspection path optimization method based on binocular vision navigation according to claim 1 is characterized in that: The key feature points of a device include the edges and corners of the device.

6. The substation inspection path optimization method based on binocular vision navigation according to claim 1 is characterized in that: The DIJKSTRA algorithm is used to determine the shortest inspection path planning.

Images