Mining conveyor belt defect detection method based on stereoscopic vision

Abstract

The invention relates to the technical field of visual inspection, and discloses a mining conveyor belt defect detection method based on stereoscopic vision. According to the mining conveyor belt defect detection method based on stereoscopic vision, a high-resolution industrial camera and an advanced image processing technology are utilized, real-time monitoring and automatic detection of the conveyor belt can be achieved, various defects can be recognized and positioned in time through image preprocessing, feature extraction and application of a machine learning model, and the defect detection efficiency is improved. According to the method, the problem is processed at the initial stage of occurrence, production stagnation and equipment damage caused by defects are avoided, meanwhile, a detailed defect detection report is generated, and a detection result is displayed in a visual form, so that an operator can quickly know the state of the equipment and make a response, and the working efficiency is improved. According to the method, the production safety and the management level are effectively improved, the efficiency and the accuracy of defect detection are remarkably improved, the dependence on manual inspection is reduced, and the risk of human errors is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of visual detection, and specifically to a method for detecting defects in mine conveyor belts based on stereo vision. Background Art

[0002] Mine conveyor belts are essential equipment in industries such as mining, metallurgy, and construction. Their main function is to transport materials such as ores, coals, and sands from one location to another. In complex environments such as mines and smelters, due to the heavy load, large friction, and harsh working conditions that the conveyor belts bear, they are prone to wear, cracks, tears, and various other defects. These defects not only affect the normal operation of the conveyor belts but may also lead to a decrease in production efficiency, equipment damage, and even safety accidents. Therefore, it is particularly important to conduct real-time monitoring and defect detection on mine conveyor belts.

[0003] Traditional methods for detecting conveyor belt defects mainly rely on manual inspections and regular maintenance. This method is not only inefficient but also easily affected by human factors, resulting in misjudgments and production losses. Summary of the Invention

[0004] Technical Problem to be Solved Aiming at the deficiencies of the prior art, the present invention provides a method for detecting defects in mine conveyor belts based on stereo vision. By using high-resolution industrial cameras and advanced image processing technologies, it can achieve real-time monitoring and automated detection of conveyor belts. Through image preprocessing, feature extraction, and the application of machine learning models, it can promptly identify and locate various defects, ensuring that problems are addressed at the initial stage, avoiding production stagnation and equipment damage caused by defects. At the same time, by generating detailed defect detection reports and presenting the detection results in a visual form, operators can quickly understand the status of the equipment and make responses, effectively improving production safety and management levels, significantly enhancing the efficiency and accuracy of defect detection, reducing the reliance on manual inspections, and lowering the risk of human errors.

[0005] (II) Technical Solution To achieve the above object, the present invention provides the following technical solution: A method for detecting defects in mine conveyor belts based on stereo vision, comprising the following steps: S1. Use a high-resolution industrial camera to collect images of the mine conveyor belt; S2. Preprocess the collected images of the mine conveyor belt, including image denoising, enhancing image contrast, enhancing image brightness, and image grayscale conversion; S3. According to the preprocessed images of the mine conveyor belt, extract the edge features, texture features, and shape features of the conveyor belt, and calculate the feature means and histograms for preliminary identification of abnormal situations of the conveyor belt; S4. Train a defect classification model through a machine learning algorithm, and input the extracted features into the trained model for defect identification and location; S5. Visualize the conveyor belt defect identification results graphically, mark the defect positions, and generate a defect detection report to output to the industrial monitoring large screen.

[0006] Preferably, the formula for image denoising is as follows: , in the formula, represents the denoised image at position, , represents the coordinates of the filter center, represents the width controlling the Gaussian distribution, represents the base of the natural logarithm, which is 2.718, represents the pi, which is 3.14.

[0007] Preferably, the formula for enhancing image contrast is as follows: , in the formula, represents the image after enhancing the contrast, represents the pixel value of the original image at point , represents the contrast gain factor, represents the brightness offset value.

[0008] Preferably, the formula for enhancing image brightness is as follows: , in the formula, represents the image after enhancing the brightness, represents the image after enhancing the contrast, represents the increased amount of brightness.

[0009] Preferably, the formula for image grayscale conversion is as follows: , in the formula, represents the pixel value of the grayscale image at point , represents the red channel value of the original conveyor belt color image at position , represents the green channel value of the original conveyor belt color image at position , represents the blue channel value of the original conveyor belt color image at position .

[0010] Preferably, the formula for extracting the edge features of the conveyor belt is as follows: , in the formula, represents the edge intensity, , respectively represent the horizontal and vertical gradients of the conveyor belt image.

[0011] Preferably, the formula for extracting the texture features of the conveyor belt is as follows: , in the formula, represents the number of occurrences of a specific gray value pair in the gray level co-occurrence matrix, represents the gray value of the conveyor belt image at position , represents that when the pixel value in the image is and returns 1, otherwise returns 0.

[0012] Preferably, the formula for extracting the shape features of the conveyor belt is as follows: , in the formula, represents the shape feature value of the conveyor belt, , represents the coordinates of the th point on the contour, represents the number of pixel points of the contour.

[0013] Preferably, the formula for calculating the feature mean is as follows: , in the formula, represents the feature mean, represents the number of feature values, represents the th feature value, represents the index subscript.

[0014] Preferably, the calculation formula of the histogram is as follows: , in the formula, represents the number of occurrences of pixels with the value of in the image, represents the original image, represents the indicator function, which returns 1 when and returns 0 otherwise.

[0015] Compared with the prior art, the present invention provides a method for detecting defects of a mine conveyor belt based on stereo vision, which has the following beneficial effects: 1. The present invention can realize real-time monitoring of the conveyor belt, significantly improve the timeliness of defect detection, and reduce production stagnation and losses caused by defects.

[0016] 2. The present invention reduces the dependence on manual inspections through an automated detection process, reduces the risk of human errors, and thus improves the reliability and consistency of the detection results.

[0017] 3. By generating detailed defect detection reports and visual results, the present invention enables operators to intuitively understand the equipment status, facilitating quick decision-making and subsequent maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of the method steps of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] 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.

[0020] Aiming at the problem that the traditional conveyor belt defect detection method mainly relies on manual inspections and regular maintenance, which is not only inefficient but also easily affected by human factors, resulting in misjudgments and production losses, a stereovision-based mine conveyor belt defect detection method is proposed. Please refer to Figure 1 , and the method includes the following steps: S1. Use a high-resolution industrial camera to collect images of the mine conveyor belt; this high-resolution industrial camera is equipped with an advanced optical lens and a high-sensitivity image sensor, which can capture fine details and changes in a complex mine environment. The acquisition process uses appropriate lighting techniques, such as LED ring lights or point light sources, to ensure the clarity and contrast of the images under different lighting conditions. At the same time, the industrial camera may incorporate high-speed shutter and dynamic range extension technologies to prevent blurring and noise caused by conveyor belt movement or environmental factors. After image acquisition, the system can transmit data to the central processing unit in real time for subsequent image processing and analysis, providing guarantee for real-time monitoring of the safety and effectiveness of the conveyor belt. This high-precision and high-efficiency image acquisition method significantly improves the accuracy and response speed of defect detection; S2. Preprocess the collected images of the mine conveyor belt, including image denoising, enhancing image contrast, enhancing image brightness, and image grayscale conversion; first, through image denoising, use a Gaussian filter for smoothing processing, and its formula is: , this step effectively reduces the random noise in the image, thus preserving important edges and details, improving the quality of subsequent feature extraction and detection. After denoising, the readability of the image is greatly improved, and the interference of noise on defect judgment is avoided; Next, by enhancing the image contrast, using the linear transformation formula: , enhancing the image contrast is an important method to improve the difference between target features and the background. By using contrast enhancement technology, the gap between different gray values can be made more obvious, thus highlighting the difference between subtle defects and normal areas. Especially in the images of mine conveyor belts, due to the possible complex textures and backgrounds, good contrast can help the algorithm quickly locate the position of defects, reducing the detection error caused by background interference. In addition, enhancing the contrast can also improve the visual readability of the image, enabling potential problems to be discovered more quickly during manual review and improving work efficiency; Subsequently, during the process of enhancing the image brightness, the formula: , by adding an appropriate constant , the overall brightness of the image can be effectively increased, ensuring that important features are not obscured due to insufficient light under different ambient lighting conditions. Such brightness enhancement is particularly crucial in the mine environment because changes in the ambient light source may cause some defects to go undetected in a timely manner; Finally, the operation of image grayscale uses the formula: , converting the color image into a grayscale image. This process not only simplifies the complexity of data processing but also reduces the consumption of computing resources. The grayscale image can effectively eliminate unnecessary color information and focus more on the analysis of shape and texture features, further improving the efficiency and accuracy of defect detection; Through these preprocessing steps, the quality of the image can be significantly improved, the influence of background noise can be reduced, and the contrast between target features and the background can be enhanced, thus laying a solid foundation for subsequent feature extraction and defect recognition. The combined effect of this series of processing means makes the system more robust and reliable when performing defect detection, able to detect potential problems in a timely manner and ensure the safe operation of the mine conveyor belt; S3. According to the preprocessed image of the mine conveyor belt, extract the edge features, texture features, and shape features of the conveyor belt, and calculate the feature mean and histogram for the preliminary identification of abnormal conditions of the conveyor belt; First, the extraction of edge features is through the following formula: , the extraction of edge features helps to identify the boundaries and direction changes of the conveyor belt, which is crucial for detecting the positions of cracks, tears, and other defects. Through precise edge detection, the system can distinguish abnormal areas from normal areas to improve the reliability of subsequent analysis; Secondly, the extraction of texture features is often achieved through the Gray-Level Co-Occurrence Matrix (GLCM), and the formula is as follows: , this method counts the frequency of gray value pairs appearing in the image, can effectively capture the subtle texture changes on the conveyor belt surface, identify texture abnormalities caused by wear, corrosion or other factors. Texture analysis not only helps to detect visual defects, but also reveals potential problems in material properties, thus assisting in judging the overall health condition of the conveyor belt; In addition, the shape feature extraction is through the following formula: , the shape feature can provide information about the defect morphology, such as the length and location of cracks, which has important guiding significance for subsequent maintenance and repair decisions. Accurate shape analysis can insight into the severity of the defect and its possible development trend; Next, it is also necessary to quantify the extracted features. The formula used for calculating the feature mean is: , where is the extracted feature value. This calculation helps to understand the overall distribution of the features and provides a reference for identifying potential abnormal patterns; Finally, the formula for histogram calculation is: , the histogram can show the distribution of different gray values in the image, thus reflecting the contrast and brightness state of the image to a certain extent. By analyzing the histogram features, the system can identify potential abnormalities in the image and then conduct deeper analysis; by extracting edge features, texture features and shape features from the preprocessed image, combined with the calculation of feature mean and histogram, not only can comprehensively understand the current state of the conveyor belt, but also can effectively warn of potential faults. This multi-dimensional feature extraction and analysis method makes defect detection more accurate and reliable, providing a strong guarantee for the safe operation of mine conveyor belts; S4. Train a defect classification model through machine learning algorithms, and input the extracted features into the trained model for defect identification and location. Training a defect classification model through machine learning algorithms is a key step in the defect detection process. First, the extracted features (such as edge features, texture features, and shape features) will be used as input data. Using the method of supervised learning, it is trained by a dataset labeled with defective and normal samples. During the training process, various machine learning algorithms can be used, such as support vector machine (SVM), decision tree, or more complex deep learning models (such as convolutional neural network CNN). These algorithms are used to obtain complex relationships and patterns between features. The goal during the training process is to enhance the generalization ability of the model so that it can accurately identify defects on new data that has not been seen before. Once the training is completed, the model will be applied to newly collected image data. Input the extracted features into the trained model, and the model will perform defect identification and location. This process can not only identify potential defects on the conveyor belt, such as cracks, wear, or tears, but also provide the precise location and type of the defects. This process usually comes with corresponding confidence scores to guide engineers to prioritize the areas that are the most serious or most likely to cause failures during maintenance and repair. S5. Visualize the conveyor belt defect identification results graphically, mark the defect locations, and generate a defect detection report to output to the industrial monitoring large screen. After defect identification and location are completed, the system will visualize the results graphically, including marking the identified defect locations on the original image. This visualization process is usually achieved with the help of powerful image processing software, which can display the type, location, and severity level of the defects on the monitoring interface in real time, enabling operators to quickly identify the problem areas. In addition, the system will also generate a detailed defect detection report, summarizing the identification results and visualization graphics, and outputting them to the industrial monitoring large screen. Such a report not only provides immediate decision-making support for operators but also serves as reference materials for later maintenance to ensure the continuous and stable operation of the conveyor belt.

[0021] By combining advanced machine learning technologies with visualization tools, this process not only improves the accuracy and efficiency of defect detection but also enhances the real-time monitoring ability of operators for the health status of the conveyor belt, thus providing a solid guarantee for the safety and economy of mining operations.

[0022] 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 spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for detecting defects of a mine conveyor belt based on stereo vision, characterized in that, It includes the following steps: S1. Use a high-resolution industrial camera to collect images of the mining conveyor belt; S2. Preprocess the collected images of the mining conveyor belt, including image denoising, enhancing image contrast, enhancing image brightness, and image grayscale conversion; S3. According to the preprocessed images of the mining conveyor belt, extract the edge features, texture features, and shape features of the conveyor belt, and calculate the feature mean and histogram for preliminary identification of conveyor belt anomalies; S4. Train a defect classification model through a machine learning algorithm, and input the extracted features into the trained model for defect identification and location; S5. Visualize the conveyor belt defect identification results graphically, mark the defect locations, and generate a defect detection report for output to the industrial monitoring large screen.

2. The method for detecting defects of a mine conveyor belt based on stereo vision according to claim 1, characterized in that: The formula for the image denoising is as follows: , in the formula, represents the denoised image at position, , represents the coordinates of the filter center, represents the width of the Gaussian distribution control, represents the base of the natural logarithm, which is 2.718, represents the pi, which is 3.

14.

3. The method for detecting defects of a mine conveyor belt based on stereo vision according to claim 2, characterized in that: The formula for enhancing the image contrast is as follows: , in the formula, represents the image after enhanced contrast, represents the pixel value of the original image at point , represents the contrast gain factor, represents the brightness offset value.

4. The method for detecting defects of a mine conveyor belt based on stereo vision according to claim 3, characterized in that: The formula for enhancing the image brightness is as follows: In the formula, represents the image after enhancing the brightness, represents the image after enhancing the contrast, represents the increased amount of brightness.

5. A method for detecting defects in a mine conveyor belt based on stereo vision according to claim 4, characterized in that: The formula for the image grayscale conversion is as follows: , in the formula, represents the pixel value of the grayscale image at point . represents the red channel value of the original conveyor belt color image at position . represents the green channel value of the original conveyor belt color image at position . represents the blue channel value of the original conveyor belt color image at position .

6. The method for detecting defects of a mine conveyor belt based on stereo vision according to claim 5, characterized in that: The formula for extracting the edge features of the conveyor belt is as follows: , in the formula, represents the edge strength, , respectively represent the horizontal and vertical gradients of the conveyor belt image.

7. A method for detecting defects in a mine conveyor belt based on stereo vision according to claim 6, characterized in that: The formula for extracting the texture features of the conveyor belt is as follows: In the formula, represents the number of occurrences of a specific gray - level value pair in the gray - level co - occurrence matrix, and represents the gray - level value of the conveyor - belt image at position . represents returning 1 when the pixel value in the image is and , and returning 0 otherwise.

8. The method for detecting defects of a mine conveyor belt based on stereo vision according to claim 7, characterized in that: The formula for extracting the shape features of the conveyor belt is as follows: In the formula, represents the shape characteristic value of the conveyor belt, , represents the coordinates of the th point on the contour, represents the number of pixel points of the contour.

9. A method for detecting defects in a mine conveyor belt based on stereo vision according to claim 8, characterized in that: The formula for calculating the feature mean is as follows: In the formula, represents the feature mean value, represents the number of feature values, represents the th eigenvalue, represents the index subscript.

10. A method for detecting defects in a mine conveyor belt based on stereo vision according to claim 9, characterized in that: The calculation formula for the histogram is as follows: In the formula, represents the number of occurrences of the pixel with the value of in the image, represents the original image, represents the indicator function, which returns 1 when and returns 0 otherwise.

Images