Pole tower defect detection system and method based on unmanned aerial vehicle inspection image

Abstract

The invention relates to the technical field of image analysis, and discloses a tower defect detection system and method based on an unmanned aerial vehicle inspection image. According to the system and the method, high-resolution image acquisition is carried out on the tower through the unmanned aerial vehicle, the image is sent to the subsequent processing module through wireless transmission, high efficiency and real-time performance of the inspection process are ensured, the image preprocessing module optimizes the image quality through denoising, contrast adjustment and other operations, a reliable basis is provided for feature extraction, and the inspection efficiency is improved. The feature extraction module further analyzes the processed image and extracts edge, texture and shape features, and then defect detection is performed through a machine learning algorithm, so that the detection accuracy is improved, the position, the type and the severity of the defect can be quickly identified, and the result evaluation module arranges the detection result into an evaluation report, so that the accuracy of defect detection is improved. And the information is timely sent to a mobile phone terminal of operation and maintenance personnel, so that the operation and maintenance personnel can quickly respond and process potential problems, and the efficiency and safety of tower inspection are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image analysis, and particularly to a tower defect detection system and method based on drone inspection images. Background Art

[0002] As an important part of the power and communication networks, towers undertake the key tasks of transmitting power and signals. With the continuous increase in power and communication demands, the maintenance and management of towers become increasingly important. During the long-term use of towers, they may be affected by various factors such as natural environment, equipment aging, and external impacts, resulting in different degrees of damage and defects, such as corrosion, deformation, fracture, etc. These defects not only affect the structural safety of the towers but may also lead to the interruption of power and communication services, thus affecting the lives and work of a wide range of users. Therefore, regular inspection of towers to timely detect and handle potential defects is an important measure to ensure the stable operation of power and communication systems.

[0003] Traditional tower inspection methods mostly rely on manual inspection, often requiring professional personnel to climb towers for on-site inspection. This not only consumes time and effort but also poses certain safety risks. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a tower defect detection system and method based on drone inspection images. High-resolution images of towers are collected by drones and transmitted to the subsequent processing module via wireless transmission, ensuring the efficiency and real-time nature of the inspection process. The image preprocessing module optimizes the image quality through operations such as denoising and contrast adjustment, providing a reliable basis for feature extraction. The feature extraction module further analyzes the processed image to extract edge, texture, and shape features, and then defect detection is carried out through machine learning algorithms, not only improving the accuracy of detection but also being able to quickly identify the location, type, and severity of defects. The result evaluation module organizes the detection results into an evaluation report and sends it to the mobile phone of the operation and maintenance personnel in a timely manner, enabling the operation and maintenance personnel to quickly respond to and handle potential problems, greatly improving the efficiency and safety of tower inspection.

[0005] To achieve the above object, the present invention provides the following technical solution: A tower defect detection system based on drone inspection images, comprising an image acquisition module, an image preprocessing module, a feature extraction module, a defect detection module, and a result evaluation module;

[0006] The image acquisition module is used to obtain high-resolution tower images taken by drone inspection and transmit them to the image preprocessing module via wireless means;

[0007] The image preprocessing module performs denoising, contrast adjustment, brightness correction, image cropping, and image scaling operations on the pole tower image, and the processed image is sent to the feature extraction module;

[0008] The feature extraction module extracts features from the processed pole tower image, extracts the pole tower edge features, pole tower texture features, and pole tower shape features and outputs them to the defect detection module;

[0009] The defect detection module performs pole tower defect detection on the pole tower edge features, pole tower texture features, and pole tower shape features through machine learning, and outputs the pole tower detection results. The pole tower detection results include the defect location, type, and severity;

[0010] The result evaluation module evaluates the pole tower detection results and outputs a pole tower defect evaluation report to be sent to the mobile phone of the pole tower operation and maintenance personnel.

[0011] Preferably, the formula for image denoising is as follows:

[0012]

[0013] In the formula, represents the pixel value of the denoised image at position , represents the pixel value of the original input image at position , represents the Gaussian kernel function, represents the normalization factor of the Gaussian kernel, , represent the relative positions around the current calculated pixel point, represents the size of the Gaussian kernel.

[0014] Preferably, the formula for contrast adjustment is as follows:

[0015]

[0016] In the formula, represents the pixel value of the adjusted image at position , represents the contrast gain factor, represents the average image brightness, represents the target brightness value, represents the pixel value of the original input image at position .

[0017] Preferably, the formula for brightness correction is as follows:

[0018]

[0019] In the formula, represents the pixel value of the corrected image at position , represents the constant for brightness correction.

[0020] Preferably, the formula for image cropping is as follows:

[0021]

[0022] In the formula, represents the pixel value of the cropped image at position , , represent the upper left coordinates of the cropping area, , represent the width and height of the cropping.

[0023] Preferably, the formula for image scaling is as follows:

[0024]

[0025] In the formula, represents the pixel value of the scaled image at position , represents the pixel value at position in the original input image, represents the weight function used for weighting during the interpolation process.

[0026] Preferably, the formula for extracting the edge features of the pole tower is as follows:

[0027]

[0028] In the formula, represents the edge intensity, and respectively represent the gradients of the pole tower image at position.

[0029] Preferably, the formula for extracting the texture features of the pole tower is as follows:

[0030]

[0031] In the formula, represents the local binary pattern calculated at , represents the pixel value at in the image, represents the position of the neighboring pixel, represents the threshold function, represents the number of pixels in the neighborhood, represents the total number of pixels.

[0032] Preferably, the formula for extracting the shape features of the pole tower is as follows:

[0033]

[0034] In the formula, represents and the central moments of order and represent the width and height of the image, represents the pixel value of the image at this position.

[0035] A method for detecting pole tower defects based on UAV inspection images includes the following steps:

[0036] S1. Obtain high-resolution pole tower images taken by UAV inspection;

[0037] S2. Perform denoising, contrast adjustment, brightness correction, image cropping, and image scaling operations on the pole tower images;

[0038] S3. Perform feature extraction based on the processed pole tower images, and extract pole tower edge features, pole tower texture features, and pole tower shape features;

[0039] S4. Use machine learning to perform pole tower defect detection on the pole tower edge features, pole tower texture features, and pole tower shape features, and output the pole tower detection results. The pole tower detection results include the defect location, type, and severity;

[0040] S5. Evaluate the pole tower detection results, and output a pole tower defect evaluation report to be sent to the mobile phone of the pole tower operation and maintenance personnel.

[0041] Compared with the prior art, the present invention provides a system and method for detecting pole tower defects based on UAV inspection images, and has the following beneficial effects:

[0042] The present invention collects high-resolution images of pole towers through UAVs, and uses wireless transmission to send the images to the subsequent processing module, ensuring the efficiency and real-time nature of the inspection process. The image preprocessing module optimizes the image quality through operations such as denoising and contrast adjustment, providing a reliable basis for feature extraction. The feature extraction module further analyzes the processed images, extracts edge, texture, and shape features, and then performs defect detection through machine learning algorithms, not only improving the accuracy of detection, but also being able to quickly identify the location, type, and severity of defects. The result evaluation module organizes the detection results into an evaluation report and sends it to the mobile phone of the operation and maintenance personnel in a timely manner, enabling the operation and maintenance personnel to quickly respond to and handle potential problems, greatly improving the efficiency and safety of pole tower inspection. Description of the Drawings

[0043] Figure 1 This is a schematic diagram of the system process of the present invention.

[0044] Figure 2 This is a schematic diagram of the method steps of the present invention. Specific embodiments

[0045] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0046] Aiming at the problem that the traditional tower inspection method mostly relies on manual inspection, and often requires professional personnel to climb the tower for on-site inspection, which is not only time-consuming and laborious, but also has certain safety risks. For this reason, a tower defect detection system based on unmanned aerial vehicle (UAV) inspection images is proposed. Please refer to Figure 1 This system includes an image acquisition module, an image preprocessing module, a feature extraction module, a defect detection module, and a result evaluation module;

[0047] The image acquisition module is a key component in the UAV inspection system and is responsible for obtaining high-resolution tower images. This module is usually equipped with a high-performance camera system, such as a digital single-lens reflex (DSLR) camera or a professional visible light camera, which can take clear images under various lighting and climatic conditions. The image acquisition module carried by the UAV can be adjusted three-dimensionally dynamically to quickly locate the target tower, so as to obtain the best shooting angle and field of view. In addition, the image acquisition module is equipped with a stabilizer and a gimbal to reduce image blurring caused by flight vibration;

[0048] Once the image capture is completed, the system will use wireless transmission technology (such as Wi-Fi or 5G network) to send the image data to the image preprocessing module in real time. This process not only ensures the instant transmission of high-resolution images, but also uses an efficient data compression algorithm to reduce the bandwidth requirement and achieve an efficient image data stream. In this way, the inspection personnel can quickly receive the images, providing strong support for subsequent image analysis and processing, and thus improving the inspection efficiency and safety;

[0049] The image preprocessing module is responsible for performing a series of processing on the obtained tower images to improve the accuracy and efficiency of subsequent feature extraction. First, the module uses an efficient denoising algorithm, such as Wiener filtering or bilateral filtering, to remove random noise in the image, thereby reducing the influence of environmental factors on the image quality. Next, contrast adjustment is performed through image enhancement technology, using methods such as histogram equalization or adaptive histogram equalization to make the important details in the image more prominent and improve the visual effect;

[0050] To ensure the consistency of images under different lighting conditions, the preprocessing module also includes a brightness correction function, which automatically adjusts the brightness of the image using linear or non-linear transformation methods to ensure its reliability during the analysis process. In addition, the image cropping function ensures that only the parts related to the tower are retained, removing redundant background information and improving the operation efficiency. Finally, through the image scaling operation, the size of the image is adjusted to an appropriate proportion to ensure that the feature extraction module can perform subsequent analysis at the best resolution;

[0051] Among them, the formula for image denoising is as follows:

[0052]

[0053] The denoising technique can effectively reduce random noise in the image, such as salt-and-pepper noise or low-contrast noise, thus ensuring that the image used for subsequent feature extraction is clearer. In the formula, represents the pixel value of the denoised image at position , represents the pixel value of the original input image at position , represents the Gaussian kernel function, represents the normalization factor of the Gaussian kernel, , represent the relative positions around the current calculated pixel point, represents the size of the Gaussian kernel. After denoising, the edges and details will be more prominent, so that feature detection algorithms (such as the Canny edge detection algorithm) can more accurately identify and locate the structural features and defects of the tower, providing more accurate basic data for subsequent analysis;

[0054] The formula for contrast adjustment is as follows:

[0055]

[0056] Through contrast adjustment, darker or brighter areas can become more prominent, which is crucial for details of the tower (such as welds, rust, etc.), making defects easier to identify. In the formula, represents the pixel value of the adjusted image at position , represents the contrast gain factor, represents the average brightness of the image, represents the target brightness value, represents the original input image at position The pixel values of images taken under different environmental conditions may have significant color differences. Through contrast adjustment, the images under different conditions can be made more consistent in terms of features, providing a reliable basis for subsequent pattern recognition and classification;

[0057] The formula for brightness correction is as follows:

[0058]

[0059] For images taken under different climate conditions or at different times, there may be brightness differences due to different lighting. Brightness correction can ensure that the brightness of the dark and bright parts of the image remains consistent, avoiding detection errors caused by environmental factors when performing defect detection. In the formula, represents the pixel value of the corrected image at position , represents the constant for brightness correction. Through brightness correction, the overall clarity and readability of the image are improved, making details and defects more obvious, which is particularly important for visual inspection;

[0060] The formula for image cropping is as follows:

[0061]

[0062] Image cropping is used to remove irrelevant backgrounds and noises, focusing the analysis on the tower and its related details. This can reduce the computational burden and improve the processing speed. In the formula, represents the pixel value of the cropped image at position , , represent the upper-left coordinates of the cropping area, , represent the width and height of the cropping. The size of the cropped image is reduced, lowering the computational complexity and memory requirements for subsequent image processing, making feature extraction and defect detection more efficient;

[0063] The formula for image scaling is as follows:

[0064]

[0065] Different algorithms and models may have specific requirements for the size of the input image. Image scaling ensures that the input meets these standards, thereby improving the compatibility and effectiveness of the algorithm. In the formula, represents the pixel value of the scaled image at position , represents the pixel value at position in the original input image, It represents the weight function used for weighting in the interpolation process. By properly scaling, the size of the image file can be reduced, thereby improving the storage and transmission efficiency of the image, especially in the UAV data monitoring system that requires real-time transmission;

[0066] The feature extraction module is the core component of the UAV inspection system. It is responsible for in-depth analysis of the tower image after image preprocessing to extract key features. This process usually involves multiple advanced image processing technologies;

[0067] First, the module extracts the edge features of the tower through edge detection algorithms (such as Canny edge detection or Sobel operator). These edge features include not only the outline of the tower, but also important structural parts such as joints, welds and fusion points, thus providing an accurate basis for the location and identification of defects. Next, the module uses texture analysis techniques, such as gray-level co-occurrence matrix (GLCM) and local binary pattern (LBP), to extract the texture features of the tower. These features are used to describe subtle changes in the surface, such as rust and coating wear. This helps to significantly improve the sensitivity and accuracy of defect detection, and also provides rich information for subsequent status assessment.

[0068] In addition, through contour analysis, the shape feature extraction module can obtain the geometric properties of the tower, such as size, proportion and curvature. These shape features are particularly important for detecting structural deformation, deviation or damage of the tower, and can help the system automatically identify abnormal conditions and improve the intelligence and automation level of inspection.

[0069] The formula for extracting tower edge features is as follows:

[0070]

[0071] Edge features are an important basis in image processing. Clear edges enable the detection algorithm to more effectively identify the outlines, joints and other structural features of the tower, and to promptly detect potential defects and damages. In the formula, represents the edge strength, and Respectively represent the tower image in The gradient of the position, through precise edge features, subsequent analysis such as morphological processing and defect classification will become more efficient, ensuring that evaluation and decision-making can be carried out quickly and accurately in complex environments;

[0072] The formula for extracting tower texture features is as follows:

[0073]

[0074] Texture features can effectively reflect the details of the image, such as problems like rust and corrosion, which are crucial for evaluating the structural integrity and safety of the pole tower. In the formula, represents the local binary pattern calculated at , represents the pixel value at in the image, represents the neighborhood pixel positions, represents the threshold function, represents the number of pixels in the neighborhood, represents the total number of pixels. Texture features can be analyzed at multiple scales and directions, which helps to more comprehensively understand the state of the pole tower surface and its potential local defects;

[0075] The formula for extracting the shape features of the pole tower is as follows:

[0076]

[0077] Shape feature extraction can clearly identify the geometric structure of the pole tower and detect any deformation or damage caused by the environment or time. In the formula, represents and order central moments, , represent the width and height of the image, represents the pixel value at this position in the image. Combining the extracted shape features, the system can provide accurate decision-making support for the operation and maintenance personnel on when to perform maintenance or updates, thus better ensuring the safe operation of the pole tower;

[0078] The defect detection module is a crucial part of the UAV inspection system, responsible for using advanced machine learning techniques to deeply analyze the extracted edge features, texture features, and shape features of the pole tower to achieve effective defect detection. This module first uses a trained model (such as support vector machine, random forest, or deep convolutional neural network) to classify and discriminate the features. By learning the features of a large number of labeled samples and comparing them with the features in the current pole tower image, the module can accurately identify possible defects, including rust, cracks, misalignment, fractures, etc. The detection results not only feedback the location and type of the defects but also evaluate their severity according to the preset standards. This process ensures the high accuracy and fast response ability of defect detection, providing a scientific basis for subsequent maintenance decisions;

[0079] To further improve the intelligence level of the system, the result evaluation module will comprehensively evaluate the pole and tower detection results. By using data analysis and visualization technologies, the detection results will be sorted out and summarized to generate a detailed pole and tower defect evaluation report. This report includes the specific location, type, and severity of the defects, and also attaches corresponding treatment suggestions and priority prompts, which is convenient for the operation and maintenance personnel to arrange according to the actual situation. Finally, this evaluation report will be sent to the mobile phones of the pole and tower operation and maintenance personnel in a timely manner through reliable wireless communication methods (such as text messages, application push, or emails). In this way, the operation and maintenance personnel can receive important inspection information in the first time, achieving rapid response and efficient management, ensuring the safe operation and timely maintenance of the pole and tower. This intelligent process effectively improves the efficiency and accuracy of the inspection work, and helps to enhance the operational safety and reliability of the pole and tower.

[0080] Please refer to Figure 2 , a method for detecting pole and tower defects based on UAV inspection images, comprising the following steps:

[0081] S1. Obtain high-resolution pole and tower images taken by UAV inspection;

[0082] S2. Perform denoising, contrast adjustment, brightness correction, image cropping, and image scaling operations on the pole and tower images;

[0083] S3. Extract features from the processed pole and tower images, including pole and tower edge features, pole and tower texture features, and pole and tower shape features;

[0084] S4. Use machine learning to detect pole and tower defects based on the pole and tower edge features, pole and tower texture features, and pole and tower shape features, and output the pole and tower detection results, which include the defect location, type, and severity;

[0085] S5. Evaluate the pole and tower detection results and output a pole and tower defect evaluation report to be sent to the mobile phones of the pole and tower operation and maintenance personnel.

[0086] Through the comprehensive application of the above methods and systems, the efficiency and safety of pole and tower inspection have been greatly improved.

[0087] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirits of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A tower defect detection system based on drone inspection images, characterized by: It includes an image acquisition module, an image preprocessing module, a feature extraction module, a defect detection module and a result evaluation module; the image acquisition module is used to obtain high-resolution tower images taken by drone inspection and transmit them to the image preprocessing module wirelessly; the image preprocessing module performs denoising, contrast adjustment, brightness correction, image cropping and image scaling operations on the tower images, and the processed images are sent to the feature extraction module; The feature extraction module performs feature extraction based on the processed tower image, extracts the tower edge features, tower texture features and tower shape features and outputs them to the defect detection module; the defect detection module performs tower defect detection on the tower edge features, tower texture features and tower shape features through machine learning, and outputs the tower detection results, which include the defect location, type and severity; the result evaluation module evaluates the tower detection results, and outputs the tower defect evaluation report and sends it to the mobile phone of the tower operation and maintenance personnel.

2. The tower defect detection system based on drone inspection images according to claim 1 is characterized by: The formula for image denoising is as follows: , in the formula, Indicates that the denoised image is at position The pixel value of Represents the original input image at position The pixel value of represents the Gaussian kernel function, represents the normalization factor of the Gaussian kernel, , Indicates the relative position around the current calculated pixel. Represents the size of the Gaussian kernel.

3. The tower defect detection system based on drone inspection images according to claim 2 is characterized by: The formula for contrast adjustment is as follows: , in the formula, Indicates that the adjusted image is at position The pixel value of represents the contrast gain factor, represents the average brightness of the image, Indicates the target brightness value, Represents the original input image at position The pixel value of .

4. The tower defect detection system based on drone inspection images according to claim 3 is characterized by: The formula for brightness correction is as follows: , in the formula, Indicates that the corrected image is at position The pixel value of A constant representing brightness correction.

5. The tower defect detection system based on drone inspection images according to claim 4 is characterized by: The formula for image cropping is as follows: , in the formula, Indicates that the cropped image is at position The pixel value of , Indicates the coordinates of the upper left corner of the clipping area. , Indicates the crop width and height.

6. The tower defect detection system based on drone inspection images according to claim 5 is characterized by: The formula for image scaling is as follows: , in the formula, Indicates that the scaled image is at position The pixel value of Represents the position in the original input image The pixel value of Represents the weight function used for weighting during interpolation.

7. The tower defect detection system based on drone inspection images according to claim 6 is characterized by: The formula for extracting the tower edge features is as follows: , in the formula, represents the edge strength, and Respectively represent the tower image in The gradient of the position.

8. The tower defect detection system based on drone inspection images according to claim 7 is characterized by: The formula for extracting the tower texture features is as follows: , in the formula, Indicated in The local binary patterns calculated at Indicates the image The pixel value at represents the neighborhood pixel position, represents the threshold function, represents the number of pixels in the neighborhood, Indicates the total number of pixels.

9. The tower defect detection system based on drone inspection images according to claim 8 is characterized by: The formula for extracting the tower shape features is as follows: , in the formula, express and The central moment of order, , Indicates the width and height of the image, Represents the pixel value of the image at that position.

10. A tower defect detection method based on drone inspection images, characterized in that: The following steps are involved: S1. Obtain high-resolution tower images taken by drone inspection; S2, performing denoising, contrast adjustment, brightness correction, image cropping and image scaling operations on the tower image; S3, performing feature extraction based on the processed tower image, extracting tower edge features, tower texture features and tower shape features; S4. Use machine learning to detect pole tower defects based on pole tower edge features, pole tower texture features, and pole tower shape features, and output the pole tower detection results, which include defect locations, types, and severity levels; S5. Evaluate the pole tower inspection results, output the pole tower defect assessment report and send it to the mobile phone of the pole tower operation and maintenance personnel.

Images