Power line fault detection and positioning method applying unmanned aerial vehicle

Abstract

The invention discloses an electric power line fault detection and positioning method applying an unmanned aerial vehicle, and the method comprises the following steps: collecting an overall distribution diagram of an electric power line, and planning an unmanned aerial vehicle inspection route according to the overall distribution diagram of the electric power line; acquiring an electric power line image through the unmanned aerial vehicle according to the unmanned aerial vehicle inspection route, and analyzing the electric power line image to obtain a fault detection result; and carrying out fault point positioning based on the fault detection result and the power line image to obtain a power line fault position. According to the invention, the efficiency and accuracy of fault detection can be greatly improved, the power failure time is reduced, and the stability of power supply is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric power maintenance, and in particular relates to a method for detecting and locating electric power line faults using an unmanned aerial vehicle. Background Art

[0002] As an important part of the power system, the safe and stable operation of power lines is crucial to ensuring social and economic activities. However, due to natural factors, equipment aging, human damage and other reasons, power lines may fail, resulting in power supply interruptions and affecting people's normal life and industrial production. Therefore, timely and accurate detection and location of power line faults is of great significance for quickly restoring power supply and reducing economic losses.

[0003] Traditional methods for detecting and locating power line faults mainly rely on manual inspections, which is inefficient and difficult to implement in complex terrain or severe weather conditions. With the rapid development of drone technology, drones are widely used in power line inspections and fault detection due to their flexibility, wide coverage, and relatively low cost. However, the inspection routes of existing drone inspections are all manually controlled and cannot be inspected along the optimal inspection route. In addition, the positioning depends on the positioning module, which increases the manufacturing cost of drones. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a method for detecting and locating power line faults using an unmanned aerial vehicle to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above object, the present invention provides a method for detecting and locating power line faults using an unmanned aerial vehicle, comprising the following steps:

[0006] Collecting an overall distribution map of power lines, and planning a drone inspection route based on the overall distribution map of power lines;

[0007] Collecting power line images by using a drone according to the drone inspection route, analyzing the power line images, and obtaining fault detection results;

[0008] The fault point is located based on the fault detection result and the power line image to obtain the power line fault position.

[0009] Preferably, the method for planning a drone inspection route includes:

[0010] Collecting remote sensing images of power lines and preprocessing them to obtain preprocessed images;

[0011] splicing the preprocessed images to obtain an overall image, performing line recognition on the overall image to obtain an overall distribution map of power lines;

[0012] Plan the drone inspection route according to the overall distribution map of the power lines.

[0013] Preferably, the method for preprocessing the power line remote sensing image comprises:

[0014] Perform atmospheric correction on remote sensing images of power lines to obtain atmospheric correction images;

[0015] The atmospheric correction image is subjected to noise reduction processing to obtain a preprocessed image.

[0016] Preferably, the method for obtaining the overall image comprises:

[0017] Searching for the same line from two images to be stitched as a stitching line, dividing the images along the stitching line, and obtaining the image to be stitched;

[0018] Two large-area images to be stitched are stitched together to obtain an initial stitched image, and a stitching seam in the initial stitched image is eliminated by a light and color uniformity method.

[0019] Preferably, the method for planning a drone inspection route includes:

[0020] Initialize algorithm parameters and model the inspection route problem as a graph search problem;

[0021] Ants are randomly placed on the graph, and each ant chooses a path based on its own pheromone concentration and heuristic factors, where the heuristic factors are distance and time;

[0022] The pheromone concentration is updated and iterated to obtain the optimal solution for the inspection path. The optimal solution for the inspection path is post-processed to obtain the optimal inspection route for the drone.

[0023] Preferably, the method for obtaining the fault detection result includes:

[0024] Building an initial fault detection model based on a convolutional neural network;

[0025] Collecting historical line fault images, training and verifying an initial fault detection model through the historical line fault images, and obtaining a fault detection model;

[0026] The power line image is identified by the fault detection model to obtain a fault detection result.

[0027] Preferably, the method for locating a fault point includes:

[0028] Extract features from the power line image containing the fault and obtain n local features;

[0029] Associating the n local features with respect to their feature positions to obtain global features of the power line image;

[0030] The features of the whole image are extracted and marked on the whole image. The global features are compared with the features on the whole image to obtain the position of the global features on the whole image, thereby obtaining the position of the fault point.

[0031] Preferably, the method for extracting features includes obtaining road morphological features and landmark shape features by edge extraction;

[0032] The edge extraction methods include: using the Sobel operator to calculate the pixel gradient;

[0033] The non-maximum pixel gradient suppression method is used to eliminate the stray effect;

[0034] Eliminate small gradient modulus points through threshold hysteresis processing;

[0035] By isolating weak edges and suppressing them, edge points are screened out to obtain edge point sets of roads and landmarks.

[0036] Compared with the prior art, the present invention has the following advantages and technical effects:

[0037] The present invention discloses a method for detecting and locating power line faults using drones, comprising the following steps: collecting an overall distribution map of the power line, planning a drone inspection route according to the overall distribution map of the power line; collecting power line images through drones according to the drone inspection route, analyzing the power line images, and obtaining fault detection results; locating the fault point based on the fault detection results and the power line images, and obtaining the power line fault position. The present invention can find the optimal inspection route through an ant colony algorithm, and can accurately locate the fault by comparing the features of the captured image with the overall image, greatly improving the efficiency and accuracy of fault detection, reducing power outage time, and ensuring the stability of power supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0039] Figure 1 The figure is a flow chart of a method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0041] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0042] Embodiment 1

[0043] like Figure 1 As shown, this embodiment provides a method for detecting and locating power line faults using a drone, comprising the following steps:

[0044] Collecting an overall distribution map of power lines, and planning a drone inspection route based on the overall distribution map of power lines;

[0045] Collecting power line images by using a drone according to the drone inspection route, analyzing the power line images, and obtaining fault detection results;

[0046] The fault point is located based on the fault detection result and the power line image to obtain the power line fault position.

[0047] Furthermore, the method for planning a drone inspection route includes:

[0048] Collecting remote sensing images of power lines and preprocessing them to obtain preprocessed images;

[0049] splicing the preprocessed images to obtain an overall image, performing line recognition on the overall image to obtain an overall distribution map of power lines;

[0050] Plan the drone inspection route according to the overall distribution map of the power lines.

[0051] Furthermore, the method for performing line recognition on the whole image based on Hough transform includes:

[0052] The preprocessed image is grayed out, and then an edge detection operator (such as the Canny operator) is used to extract edge information in the image. The purpose of this step is to identify edge pixels that may belong to power lines.

[0053] Hough transform process:

[0054] Define parameter space: Hough transform maps a straight line in image space to parameter space. In a two-dimensional image, a straight line can be represented by the polar coordinate equation ρ = xcosθ + ysinθ, where ρ is the distance from the origin to the line, and θ is the angle between the normal of the line and the x-axis.

[0055] Accumulator array initialization: Create a two-dimensional array (accumulator) to store the number of votes for each pair (ρ, θ) in the parameter space.

[0056] Voting process: For each edge point in the image space, calculate its corresponding (ρ, θ) pair and increase the vote at the corresponding position in the accumulator array.

[0057] Threshold determination: A threshold is set based on the image data to determine which (ρ, θ) pairs represent true straight lines. (ρ, θ) pairs exceeding the threshold are considered to be detected straight lines.

[0058] Line extraction process:

[0059] Based on the number of votes in the accumulator array, the (ρ,θ) pair representing the line is extracted.

[0060] These parameters are converted back to image space and the detected lines are plotted.

[0061] Post-processing process:

[0062] Straight-line connection: If the power lines are detected as multiple line segments, they can be connected through a minimum spanning tree.

[0063] False detection removal: Remove straight lines that do not meet the characteristics of power lines, such as lines that are too short or lines that do not match the known direction of power lines.

[0064] Result optimization: Further optimization steps may be required, such as line thinning, merging adjacent lines, etc.

[0065] Result output:

[0066] Output the final power line identification results.

[0067] Furthermore, the method for preprocessing the power line remote sensing image includes:

[0068] Perform atmospheric correction on remote sensing images of power lines to obtain atmospheric correction images;

[0069] The atmospheric correction image is subjected to noise reduction processing to obtain a preprocessed image.

[0070] Furthermore, the method for obtaining the overall image includes:

[0071] Searching for the same line from two images to be stitched as a stitching line, dividing the images along the stitching line, and obtaining the image to be stitched;

[0072] Two large-area images to be stitched are stitched together to obtain an initial stitched image, and a stitching seam in the initial stitched image is eliminated by a light and color uniformity method.

[0073] Furthermore, the method for planning the inspection route of the drone includes:

[0074] First, the parameters of the ant colony algorithm are initialized, and the inspection route problem is modeled as a graph search problem. The inspection route problem is to specify the starting position in the overall distribution map of the power lines, and plan the path based on the starting position and the overall distribution of the power lines. In this process, the problem is converted into a graph structure, where nodes represent inspection points and edges represent paths between inspection points. Next, ants are randomly placed on the graph, and each ant selects a path based on its own pheromone concentration and heuristic factors. The heuristic factor can be set according to actual needs, such as distance, time, etc.

[0075] When ants choose a path, they evaluate the quality of each path based on pheromone concentration and heuristic factors. Pheromone concentration is a chemical released by ants when they are looking for food, which is used to guide other ants to find the food source. In this problem, pheromone concentration can be regarded as a quality indicator of the path, and a higher pheromone concentration indicates a better path. Heuristic factors can help ants consider other factors such as distance and time when choosing a path.

[0076] After the ants choose a path, they need to update the pheromone concentration. This process includes two parts: deposition and volatilization. Deposition means that the ants leave a certain amount of pheromone on the path they have walked to enhance the attractiveness of the path. Volatilization means that the pheromone will gradually decrease over time, which can prevent the pheromone concentration on some paths from being too high, causing the ants to fall into a local optimal solution.

[0077] Through multiple iterations, the optimal solution can be found. In each iteration, the ant will select a path based on the current pheromone concentration and heuristic factor, and then update the pheromone concentration. This process will be repeated until the optimal solution that meets the requirements is found or the preset number of iterations is reached.

[0078] Finally, the optimal path needs to be post-processed and output. Post-processing can include removing redundant inspection points, optimizing the path shape, etc. Outputting the optimal inspection route is to present the final result to the user or other systems.

[0079] Furthermore, the method for obtaining the fault detection result includes:

[0080] Building an initial fault detection model based on a convolutional neural network;

[0081] Collecting historical line fault images, training and verifying an initial fault detection model through the historical line fault images, and obtaining a fault detection model;

[0082] The power line image is identified by the fault detection model to obtain a fault detection result.

[0083] Furthermore, the method for locating a fault point includes:

[0084] First, feature extraction is performed on the image of the power line containing the fault. This step involves using advanced image processing algorithms and techniques, such as edge detection technology, to identify local features in the image. Through these methods, n key local features can be extracted from the image.

[0085] Next, these local features are associated with their feature positions. This means that the n local features extracted are integrated and compared to form a more comprehensive view. The purpose of this step is to obtain the global features of the power line image, thereby providing more accurate information for subsequent fault location.

[0086] After obtaining the global features, the whole image is subjected to feature extraction. The same technique as the local feature extraction is used to ensure that the extracted features are the same. These extracted features are then annotated on the whole image for visual analysis.

[0087] Finally, the previously obtained global features are compared with the features on the overall image. This step is to find the exact location of the global features on the overall image. Through this comparison, the specific location of the fault point in the power line image can be clearly identified.

[0088] Furthermore, the method for extracting features includes obtaining road morphological features and landmark shape features by edge extraction;

[0089] Edge extraction methods include:

[0090] 1. Use the Sobel operator to calculate the pixel gradient. The operator is two 3*3 matrices, which are used to calculate the pixel gradient matrix in two directions of the image. The form is as follows:

[0091]

[0092] Where I is the image pixel matrix after the image is grayscale processed, and the * inside is regarded as a cross-correlation operation.

[0093] because

[0094] So we can get the gradient strength matrix G xy .

[0095] 2. In order to eliminate the stray effects caused by edge detection, non-maximum pixel gradient suppression came into being. Since the previously calculated gradient edge is multiple pixels wide, it needs to be optimized to an accurate point width. Therefore, it is necessary to compare the gradient strength of the current image pixel with the gradient strength of the adjacent image pixel in the positive and negative gradient directions. If the gradient strength of the current image pixel is the largest compared to the gradient strength of other pixels in the same direction, retain its value, otherwise suppress it and set it to zero. Considering the precise calculation, linear interpolation can be used between two adjacent image pixels to obtain the corresponding image pixel gradient.

[0096] 3. After shortening the edge width, there are still a large number of small gradient modulus points left, which require threshold hysteresis processing. Among them, a high threshold and a low threshold need to be set to distinguish edge pixels. After the edge pixels are screened, there will be three situations: (1) If the gradient value of the edge pixel is greater than the high threshold, it is considered a strong edge point; (2) If the gradient value of the edge pixel is between the low threshold and the high threshold, it is considered a weak edge point; (3) If the gradient value of the edge pixel is less than the low threshold, it is directly set to zero and deleted.

[0097] 4. Isolated weak edge suppression, the main purpose is to filter out qualified edge points to make the boundary more reliable. Based on the previous, strong edge points can be considered as true edges, while some weak edge points may be caused by noise changes and need to be deleted directly. It is usually believed that the edge points caused by true edges are the same as strong edge points, while the weak edge points caused by noise are not. Therefore, the depth-first algorithm can be used to check the 8-connected area pixels of each weak edge point. If there is a strong edge point next to it, then keep this weak edge point, otherwise delete it directly.

[0098] After completing the above steps, the edge point set of roads and landmarks can be obtained.

[0099] The straight lines of power lines and roads are distinguished by width thresholds, and edge shapes with prominent features are selected as landmarks, and positioning is performed based on the landmarks and road shapes.

[0100] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for detecting and locating power line faults using an unmanned aerial vehicle, characterized in that: The following steps are involved: Collecting an overall distribution map of power lines, and planning a drone inspection route based on the overall distribution map of power lines; Collecting power line images by using a drone according to the drone inspection route, analyzing the power line images, and obtaining fault detection results; The fault point is located based on the fault detection result and the power line image to obtain the power line fault position.

2. The method for detecting and locating power line faults using an unmanned aerial vehicle according to claim 1, characterized in that: The method for planning a drone inspection route includes: Collecting remote sensing images of power lines and preprocessing them to obtain preprocessed images; splicing the preprocessed images to obtain an overall image, performing line recognition on the overall image to obtain an overall distribution map of power lines; Plan the drone inspection route according to the overall distribution map of the power lines.

3. The method for detecting and locating power line faults using an unmanned aerial vehicle according to claim 2, characterized in that: The methods for preprocessing remote sensing images of power lines include: Perform atmospheric correction on remote sensing images of power lines to obtain atmospheric correction images; The atmospheric correction image is subjected to noise reduction processing to obtain a preprocessed image.

4. The method for detecting and locating power line faults using an unmanned aerial vehicle according to claim 2, characterized in that: The method for obtaining the overall image comprises: Searching for the same line from two images to be stitched as a stitching line, dividing the images along the stitching line, and obtaining the image to be stitched; Two large-area images to be stitched are stitched together to obtain an initial stitched image, and a stitching seam in the initial stitched image is eliminated by a light and color uniformity method.

5. The method for detecting and locating power line faults using an unmanned aerial vehicle according to claim 2, characterized in that: The method for planning a drone inspection route includes: Initialize algorithm parameters and model the inspection route problem as a graph search problem; Ants are randomly placed on the graph, and each ant chooses a path based on its own pheromone concentration and heuristic factors, where the heuristic factors are distance and time; The pheromone concentration is updated and iterated to obtain the optimal solution for the inspection path. The optimal solution for the inspection path is post-processed to obtain the optimal inspection route for the drone.

6. The method for detecting and locating power line faults using an unmanned aerial vehicle according to claim 1, characterized in that: The method for obtaining the fault detection result comprises: Building an initial fault detection model based on a convolutional neural network; Collecting historical line fault images, training and verifying an initial fault detection model through the historical line fault images, and obtaining a fault detection model; The power line image is identified by the fault detection model to obtain a fault detection result.

7. The method for detecting and locating power line faults using an unmanned aerial vehicle according to claim 1, characterized in that: The method for locating a fault point comprises: Extract features from the power line image containing the fault and obtain n local features; Associating the n local features with respect to their feature positions to obtain global features of the power line image; The feature extraction is performed on the whole image, and the extracted features are marked on the whole image. The global features are compared with the features on the whole image to obtain the position of the global features on the whole image, thereby obtaining the position of the fault point.

8. The method for detecting and locating power line faults using an unmanned aerial vehicle according to claim 7, characterized in that: The method for feature extraction includes obtaining road morphological features and landmark shape features by edge extraction; The edge extraction methods include: using the Sobel operator to calculate the pixel gradient; The non-maximum pixel gradient suppression method is used to eliminate the stray effect; Eliminate small gradient modulus points through threshold hysteresis processing; By isolating weak edges and suppressing them, edge points are screened out to obtain edge point sets of roads and landmarks.

Images