A method for detecting belt tear

Abstract

This application discloses a belt tear detection method, comprising: acquiring a gray-level distribution histogram corresponding to a reference image, and determining a preset gray-level threshold based on the gray-level distribution histogram; performing threshold segmentation, dilation, and enhancement operations on the reference image according to the preset gray-level threshold to form a template image; inputting the preprocessed training image into a training model for template training to obtain a template weight file; acquiring a test image and preprocessing it to obtain a preprocessed test image; matching the preprocessed test image with the template weight file; if the preprocessed test image is within the allowable range of the template weight file, recording the position information of the preprocessed test image; and using affine transformation to achieve the maximum fit between the preprocessed test image and the template image to obtain a transformed image; performing difference calculation on the transformed image and the template image to obtain a difference image; when the pixel gray-level value in the difference image is greater than the difference threshold, a belt tear is detected in the test image.

Description

Technical Field

[0001] This application relates to the field of machine vision technology, and in particular to a method for detecting belt tearing. Background Technology

[0002] Belt conveyors, as an indispensable part of transportation, play a vital role in grain transport, ore conveying, and smart mines, serving as a simple and practical conveyor method. However, with increasing usage time, belt wear is inevitable, and among belt damage, penetrating tears are the most severe. If a tear is not stopped promptly, it can cause long-distance, large-area tears, resulting in huge economic losses and even endangering the lives of operators. Therefore, real-time belt monitoring and the ability to quickly determine if a tear has occurred are of paramount importance for protecting operator safety and minimizing economic losses.

[0003] Existing belt tear detection methods mainly use machine vision algorithms. Traditional machine vision algorithms directly judge the acquired images, which not only consumes a lot of computer resources, but also has low detection efficiency and long detection time. Summary of the Invention

[0004] This application provides a belt tear detection method to improve belt tear detection efficiency.

[0005] To address the aforementioned technical problems, the embodiments of this application disclose the following technical solutions:

[0006] This application discloses a belt tear detection method, including: acquiring a gray-level distribution histogram corresponding to a reference image, and determining a preset gray-level threshold based on the gray-level distribution histogram;

[0007] Based on the preset grayscale threshold, threshold segmentation, dilation, and enhancement operations are performed on the reference image to form a template image;

[0008] The preprocessed training image is input into the training model for template training to obtain the template weight file;

[0009] Acquire the image to be tested and preprocess the image to be tested to obtain a preprocessed image to be tested;

[0010] The preprocessed test image is matched with the template weight file. If the preprocessed test image is within the allowable range of the template weight file, the position information of the preprocessed test image is recorded. Then, based on the position information of the preprocessed test image, affine transformation is used to make the preprocessed test image and the template image achieve the maximum fit, and the transformed image is obtained.

[0011] The transformed image and the template image are compared by a difference calculation to obtain a difference image. When the pixel grayscale value in the difference image is greater than the difference threshold, the belt in the image to be tested is torn.

[0012] In some embodiments, the process of obtaining the preprocessed training image includes: selecting multiple untorn and clear belt images as training images, and performing threshold segmentation, enhancement, and dilation processing on the training images according to a preset grayscale threshold to obtain the preprocessed training image.

[0013] In some embodiments, the template weight file includes feature information such as pixel grayscale values, line widths, and line areas of the laser line regions in the training image.

[0014] In some embodiments, obtaining the grayscale distribution histogram corresponding to the reference image and determining the preset grayscale threshold based on the grayscale distribution histogram includes: converting the reference image to grayscale to obtain a grayscale image of the belt surface;

[0015] Based on the grayscale image of the belt surface, the corresponding grayscale distribution histogram is obtained;

[0016] A preset grayscale threshold is determined based on the grayscale values ​​of different regions on the grayscale distribution histogram.

[0017] In some embodiments, the step of inputting the preprocessed training image into the training model for template training further includes: calculating the correlation between the preprocessed training image and the template image; when the correlation is greater than or equal to 0.9, the preprocessed training image is input into the training model for template training.

[0018] In some embodiments, threshold segmentation, dilation, and enhancement operations are performed on the reference image according to the preset grayscale threshold to form a template image, including:

[0019] A set of pixels in the reference image that fall within the preset grayscale threshold range is obtained to form an initial template region image;

[0020] The initial template region image that was ignored is dilated to form a dilated template region.

[0021] The dilated template region is cropped from the original image, and then enhanced to form a template image.

[0022] In some embodiments, the formula for the enhancement process includes:

[0023] G' = G Exponen / (1)

[0024] Where G is the grayscale value of the image before enhancement, G′ is the grayscale value of the image after exponential transformation, and Exponent is the exponent of the transformation.

[0025] In some embodiments, the step of acquiring the image to be tested and preprocessing the image to be tested to obtain a preprocessed image to be tested includes: acquiring the image to be tested and converting the image to be tested into a grayscale image to be tested;

[0026] Obtain a set of pixels in the grayscale image to be tested that meet the preset grayscale threshold range to form an initial test area;

[0027] The initial test area is dilated and enhanced to obtain a preprocessed test image.

[0028] In some embodiments, the difference calculation formula includes:

[0029] D(i,j)=|S(i,j)-T(i,j)| (2)

[0030] Wherein, S(i,j) is the gray value corresponding to the pixel at position coordinate (i,j) in the template image, T(i,j) is the gray value corresponding to the pixel at position coordinate (i,j) in the transformed image, and D(i,j) is the gray value of the pixel at position coordinate (i,j) in the difference image.

[0031] The beneficial effects of this application are:

[0032] This application discloses a belt tear detection method, comprising: acquiring a gray-level distribution histogram corresponding to a reference image, and determining a preset gray-level threshold based on the gray-level distribution histogram; performing threshold segmentation, dilation, and enhancement operations on the reference image according to the preset gray-level threshold to form a template image; inputting the preprocessed training image into a training model for template training to obtain a template weight file; acquiring a test image and preprocessing the test image to obtain a preprocessed test image; matching the preprocessed test image with the template weight file; if the preprocessed test image is within the allowable range of the template weight file, recording the position information of the preprocessed test image; then, based on the position information of the preprocessed test image, using affine transformation to achieve the maximum fit between the preprocessed test image and the template image to obtain a transformed image; performing difference calculation on the transformed image and the template image to obtain a difference image; when the pixel gray-level value in the difference image is greater than the difference threshold, a belt tear occurs in the test image, narrowing the image range for judgment and helping to improve judgment efficiency. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in this disclosure, the accompanying drawings used in some embodiments of this disclosure will be briefly described below. Obviously, the drawings described below are only drawings of some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings. In addition, the drawings described below can be regarded as schematic diagrams and are not intended to limit the actual size of the product, the actual flow of the method, the actual timing of the signals, etc. involved in the embodiments of this disclosure.

[0034] Figure 1 This is a schematic diagram of a belt tear monitoring system provided according to some embodiments;

[0035] Figure 2 This is a schematic diagram of a belt tear detection method provided according to some embodiments;

[0036] Figure 3 A grayscale distribution histogram provided in this application;

[0037] Figure 4 This is a comparison image before and after image enhancement processing, as exemplified in this application.

[0038] Figure 5 This is a schematic diagram illustrating a differential calculation as an example of this application. Detailed Implementation

[0039] The technical solutions in some embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments provided in this disclosure are within the scope of protection of this disclosure.

[0040] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this application means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0041] This application discloses a belt tear detection method. First, a template image is created. Then, multiple tear-free images are selected as training images for model training to obtain a weight file for the training model, including the pixel grayscale values, line width, and line area of ​​the laser line shape region in the training image. By thresholding the test image, the image of the laser line shape region is retained as a preprocessed test image and matched with the template weight file. When the preprocessed test image is within the allowable range of the template weight file, the preprocessed test image is affine transformed to make it overlap with the template image as much as possible, resulting in a transformed image. The difference between the transformed image and the template image is calculated to obtain a difference image. When the pixel grayscale value in the difference image is greater than the difference threshold, a tear occurs in the belt in the test image.

[0042] Figure 1 This is a schematic diagram of a belt tear monitoring system provided according to some embodiments. Figure 2 This is a schematic flowchart of a belt tear detection method provided according to some embodiments. The following is in conjunction with... Figure 1 and Figure 2 The method will be explained.

[0043] In some embodiments of this application, the belt tear image acquisition device is as follows: Figure 1 As shown, to achieve cyclic conveying, the belt conveyor is equipped with two belts that circulate. A line laser emitter 3 is positioned between the two reciprocating belts, emitting a laser beam towards the first belt. An image acquisition device 2 is also provided, located between the two reciprocating belts, to acquire an image of the laser beam on the first belt 1.

[0044] In some embodiments of this application, the image acquisition device is at the same vertical distance from the line laser emitter and the first belt. The image acquisition device is located to one side of the line laser emitter and is used to acquire images of the laser beam on the first belt.

[0045] The images of the belt surface acquired by the image acquisition device in this application are divided into two categories: training images and images to be tested.

[0046] The training images are belt surface images in their un-torn state, used to obtain the template matching weight file. The test images are belt surface images acquired during the detection process.

[0047] Because the belt changes position due to vibration during operation, multiple training images are needed to train the detection model during the training process.

[0048] The machine vision algorithm used in the belt tear detection method of this application specifically includes:

[0049] S100: Obtain the grayscale distribution histogram corresponding to the reference image, and determine the preset grayscale threshold based on the grayscale distribution histogram.

[0050] In this application, the grayscale distribution histogram corresponding to the reference image is obtained, and a preset grayscale threshold is determined based on the grayscale distribution histogram. Specifically, this includes: converting the reference image to grayscale to obtain a grayscale image of the belt surface; obtaining the corresponding grayscale distribution histogram based on the grayscale image of the belt surface; and determining the preset grayscale threshold based on the grayscale values ​​of different regions on the grayscale distribution histogram.

[0051] Figure 3 This application provides a grayscale distribution histogram. For example... Figure 3 As shown, the grayscale distribution histogram has a binary property: the closer the grayscale value is to 0, the darker the brightness; the closer the grayscale value is to 255, the brighter the brightness. Combining the brightness of the laser line with the brightness of the background color, it can be clearly determined that: low-brightness areas with grayscale values ​​close to 0 correspond to background or darker areas in the image, while high-brightness areas with grayscale values ​​close to 255 correspond to laser line areas in the image. Therefore, a preset grayscale threshold can be determined based on the grayscale distribution histogram; that is, the minimum grayscale value corresponding to the high-brightness area is used as the preset grayscale threshold. Of course, to further prevent the laser lines from being missed, a grayscale tolerance value is added to this minimum grayscale value. The specific calculation formula is: Preset grayscale threshold = Minimum grayscale value corresponding to the high-brightness area - Grayscale tolerance value.

[0052] In this application, histograms are used to determine the threshold. By combining images with histograms, image information is converted into numerical information in the histogram, making the threshold determination process more intuitive, rigorous, and persuasive.

[0053] The selection of the preset grayscale threshold includes: converting the reference image into a grayscale distribution histogram, then performing threshold segmentation on the grayscale image, and using the grayscale value range corresponding to the laser-covered area in the grayscale image as the preset grayscale threshold. For example... Figure 3 As shown, the horizontal axis represents grayscale value, and the vertical axis represents pixel area. Pixels with grayscale values ​​between 245 and 255 can cover the entire laser area; therefore, the preset grayscale threshold in this example is 245-255. The preset grayscale threshold can also be set based on the grayscale image corresponding to the reference image.

[0054] S200: Based on a preset grayscale threshold, perform threshold segmentation, dilation, and enhancement operations on the reference image to form a template image.

[0055] A set of pixels within a preset grayscale threshold range in the reference image is obtained to form an initial template region image. The initial template region image is then dilated to form a dilated template region, which is then cropped from the original image. The dilated template region is then enhanced to form a template image.

[0056] A clear image of the belt surface, without any tears or sharp edges, is selected as the reference image. The set of pixels in the reference image that fall within a preset grayscale threshold range forms the initial template region image. A dilation operation is performed on the initial template region image to form a dilated template region, which is then cropped from the original image. The dilated template region is then enhanced to form the template image.

[0057] The reference image is converted into a grayscale image, and the location information of pixels that meet the preset grayscale threshold range is obtained. The area where the pixels that meet the preset grayscale threshold range are located is the initial template area image.

[0058] The process of dilating the initial template region image includes: Based on the pixel position information of the initial template region image, obtaining pixels that are less than or equal to a preset distance from the initial template region image; this set constitutes the dilated pixel set. In this example, the preset distance is selected as 3.5 pixels. Therefore, the set of pixels located within a region less than or equal to 3.5 pixels from the initial template region image constitutes the dilated pixel set, and its position in the original image is the dilated template region.

[0059] The process involves enhancing the dilated template region to form a template image, including: acquiring the pixel location information and original grayscale values ​​of the dilated template region, and calculating the enhanced grayscale values ​​according to the enhancement transformation formula. The location information of the template image consists of the pixel location information of the dilated template region, and the grayscale values ​​of the template image are the enhanced grayscale values ​​of the pixels in the dilated template region.

[0060] The enhancement formula is shown in formula (1) below.

[0061] G' = G Exponent (1)

[0062] Where G is the gray value of the original image, G′ is the gray value of the image after exponential transformation, and Exponent is the exponent of the transformation. The value of Exponent is uncertain depending on the application scenario. Experiments have shown that an Exponent value of 2 yields the best results.

[0063] In practical applications, different algorithms can be used to enhance images. For example, contrast can be increased to improve the contrast between the laser line area and the background area; alternatively, the brightness of the laser line area can be reduced to make it brighter. Those skilled in the art can choose appropriate algorithms according to actual needs, and all of these fall within the scope of this application. In this application, increasing image brightness through image re-enhancement can, on the one hand, highlight the laser line portion, and on the other hand, suppress some noise, thereby improving the accuracy of subsequent belt tear detection.

[0064] Figure 4 Here is a comparison image before and after image enhancement processing, as an example of this application. Figure 4 As shown in the image, the image to the left of the arrow is the image before image enhancement, and the image to the right of the arrow is the image after image enhancement. After image enhancement, it can be seen that the surrounding dark areas are suppressed more obviously, and the white areas have been brightened.

[0065] S300: Select multiple untorn and clear images of the belt as training images. Perform threshold segmentation, enhancement, and dilation processing on the training images according to the preset grayscale threshold to obtain preprocessed training images.

[0066] The training image is cropped according to the position information of the dilated template region to obtain the training laser line region. Enhancement processing is applied to the training laser line region to obtain an enhanced training image, which makes the other dark areas of the training laser line region more clearly suppressed and the white areas brighter. The enhanced training image is then dilated so that the dilated region can cover all laser line regions, resulting in a preprocessed training image.

[0067] S400: Input the preprocessed training image into the training model for template training to obtain a template weight file. The template weight file includes feature information such as the pixel grayscale value, line width, and line area of ​​the laser line region in the training image.

[0068] Before inputting the preprocessed training image into the training model for template training, the correlation between the preprocessed training image and the template image can be calculated. The number of pixel matches between the preprocessed training image and the template image is calculated and denoted as the pixel match count. The ratio of the pixel match count to the total number of pixels in the preprocessed training image is denoted as the correlation between the preprocessed training image and the template image.

[0069] In this application, template training will be performed on preprocessed training images with a correlation degree greater than or equal to 0.9, while preprocessed training images with a correlation degree less than 0.9 will be discarded.

[0070] The preprocessed training image is input into the training model for template training, resulting in a template weight file. The template weight file includes feature information such as pixel grayscale values, line width, and line area of ​​the laser line region in the training image. The training model can be a deformation model.

[0071] S500: Performs threshold segmentation, dilation, and enhancement processing on the image to be tested to obtain a preprocessed image to be tested.

[0072] The test image is acquired and converted into a grayscale image. The location information of pixels in the grayscale image that fall within a preset grayscale threshold range is obtained. The region containing these pixels is the initial test region. The initial test region is then subjected to dilation and enhancement processing to obtain the preprocessed test image.

[0073] In this example, the process of dilating and enhancing the initial test area is consistent with the reference image method for the reference image described earlier, and will not be described in detail here.

[0074] In this example, the preset grayscale threshold range of the image under test is consistent with the preset grayscale threshold of the reference image.

[0075] S600: Match the preprocessed test image with the template weight file. If the preprocessed test image is within the allowable range of the template weight file, record the position information of the preprocessed test image. Then, based on the position information, use affine transformation to rotate or translate the preprocessed test image so that the preprocessed test image and the template image can achieve the maximum fit. The preprocessed test image after affine transformation is called the transformed image.

[0076] The affine transformation matrix is ​​shown below:

[0077]

[0078] The matrix at the back is the image matrix before transformation, and the matrix at the front is the matrix after transformation. The translation or rotation transformation of the original image is performed by controlling the values ​​of the small matrices R and T in the middle matrix.

[0079] S700: Perform difference calculation on the transformed image and the template image to obtain a difference image. When the pixel gray value in the difference image is greater than the difference threshold, it is determined that the belt in the test image has a tear.

[0080] The process of performing difference calculations between the transformed image and the template image to obtain a difference image includes: obtaining the pixel position information and corresponding grayscale values ​​of the transformed image; and performing difference calculations between the grayscale values ​​of the corresponding pixels of the transformed image and the template image to obtain a difference image.

[0081] The formula for calculating the difference is as follows:

[0082] D(i,j)=|S(i,j)-T(i,j)| (2)

[0083] Where S(i,j) is the gray value corresponding to a certain pixel in the template image, T(i,j) is the gray value corresponding to the pixel in the transformed image that is at the same position as the template image, and D(i,j) is the gray value corresponding to the pixel in the difference image that is at the same position as the template image.

[0084] In some embodiments of this application, in order to further reduce computation time and improve computation efficiency, before performing differential calculation on the transformed image and the template image, a set of pixels in the transformed image and the template image whose pixel grayscale values ​​are greater than the differential grayscale threshold is obtained, and differential calculation is performed on the pixels in the transformed image and the template image whose pixel grayscale values ​​are greater than the differential grayscale threshold.

[0085] For ease of calculation, the image's position coordinates can also be pre-divided pixel block coordinates, such as the original image based on...

[0086] Figure 5 As shown in the figure, this is a schematic diagram of a difference calculation example of this application. Figure 5 In this diagram, 'a' is the template image, 'b' is the transformed image, and 'c' is the difference image after the difference operation. The grayscale values ​​at corresponding positions are calculated by difference, and the information of the pixels with differences is output.

[0087] In summary, this application provides a method for detecting belt tears, which employs both laser line imaging and machine vision algorithms to process images. Specifically, it acquires an image of the belt surface after laser projection and uses machine vision algorithms to determine the presence or absence of laser lines, thereby assessing the extent of belt tearing. In this process, compared to traditional camera imaging, laser line imaging can record more information, thus ensuring more accurate acquisition of belt surface information, such as tear details, and consequently ensuring the accuracy of subsequent detection results.

[0088] On the other hand, in the process of determining belt surface tears using the machine vision algorithm of this application, the image is segmented using a laser line grayscale threshold, and the segmented region is dilated. A region of interest (ROI) containing the laser line is then cropped from the grayscale image of the belt surface. In the subsequent belt tear determination process, only the ROI image needs to be judged, greatly reducing the image range for judgment and improving both efficiency and accuracy. Furthermore, compared to the prior art of directly judging tears from the acquired image, this application preprocesses and enhances the image before tear determination, highlighting the laser line portion and effectively suppressing some noise while increasing the brightness of the laser line. This makes the belt tear judgment more accurate and reduces the likelihood of false positives. Furthermore, this application utilizes the grayscale threshold of the laser line in the grayscale distribution histogram to perform threshold segmentation on the belt surface image and dilates the segmented region. A template image containing the laser line is then cropped from the grayscale image of the belt surface. In the subsequent belt tear detection process, only the template image needs to be evaluated, significantly reducing the image range for judgment and improving both efficiency and accuracy. Differential calculation is used to highlight the difference between the test image and the template image; the calculation is simple and highly accurate.

[0089] In summary, the machine vision algorithm used in this application employs concise judgment statements, high accuracy, strong recognition, and wide applicability in the process of judging belt surface tears. It can also quickly identify belt tears, that is, it can respond in a timely manner when a tear occurs or in the early stage of a tear, thereby avoiding other failures of the belt in the future.

[0090] Since the above embodiments are all described in conjunction with other methods, and different embodiments have the same parts, the same or similar parts between the various embodiments in this specification can be referred to mutually. They will not be described in detail here.

[0091] It should be noted that in this specification, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a circuit structure, article, or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a circuit structure, article, or device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the circuit structure, article, or device that includes said element.

[0092] Other embodiments of this application will readily conceive of by those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this application are indicated by the content of the claims.

[0093] The embodiments described above do not constitute a limitation on the scope of protection of this application.

Claims

1. A method for detecting belt tearing, characterized in that, include: Obtain the grayscale distribution histogram corresponding to the reference image, and determine the preset grayscale threshold based on the grayscale distribution histogram; Based on the preset grayscale threshold, threshold segmentation, dilation, and enhancement operations are performed on the reference image to form a template image; The preprocessed training image is input into the training model for template training to obtain the template weight file; Acquire the image to be tested and preprocess the image to be tested to obtain a preprocessed image to be tested; The preprocessed test image is matched with the template weight file. If the preprocessed test image is within the allowable range of the template weight file, the position information of the preprocessed test image is recorded. Then, affine transformation is used to make the preprocessed test image and the template image achieve the maximum fit, and the transformed image is obtained. The transformed image and the template image are compared by a difference calculation to obtain a difference image. When the pixel grayscale value in the difference image is greater than the difference threshold, the belt in the image to be tested is torn. Based on the preset grayscale threshold, threshold segmentation, dilation, and enhancement operations are performed on the reference image to form a template image, including: A set of pixels in the reference image that fall within the preset grayscale threshold range is obtained to form an initial template region image; The initial template region image that was ignored is dilated to form a dilated template region. The dilated template region is cropped from the original image, and then enhanced to form a template image.

2. The belt tear detection method according to claim 1, characterized in that, The process of obtaining preprocessed training images includes: selecting multiple untorn and clear belt images as training images, and performing cropping, enhancement, and dilation processing on the training images according to a preset grayscale threshold to obtain preprocessed training images.

3. The belt tear detection method according to claim 1, characterized in that, The template weight file includes laser line region, line width, and line area feature information of the training image.

4. The belt tear detection method according to claim 1, characterized in that, The step of acquiring the grayscale distribution histogram corresponding to the reference image and determining the preset grayscale threshold based on the grayscale distribution histogram includes: The reference image is converted to grayscale to obtain a grayscale image of the belt surface; Based on the grayscale image of the belt surface, the corresponding grayscale distribution histogram is obtained; A preset grayscale threshold is determined based on the grayscale values ​​of different regions on the grayscale distribution histogram.

5. The belt tear detection method according to claim 1, characterized in that, The step of inputting the preprocessed training image into the training model for template training also includes: The correlation between the preprocessed training image and the template image is calculated. When the correlation is greater than or equal to 0.9, the preprocessed training image is input into the training model for template training.

6. The belt tear detection method according to claim 1, characterized in that, The formula for the enhancement process includes: G'=G Exponent (1) Where G is the grayscale value of the image before enhancement, G′ is the grayscale value of the image after exponential transformation, and Exponent is the exponent of the transformation.

7. The belt tear detection method according to claim 1, characterized in that, The process of acquiring the image to be tested and preprocessing the image to obtain a preprocessed image to be tested includes: Acquire the image to be tested, and convert the image to be tested into a grayscale image to be tested; Obtain a set of pixels in the grayscale image to be tested that meet the preset grayscale threshold range to form an initial test area; The initial test area is dilated and enhanced to obtain a preprocessed test image.

8. The belt tear detection method according to claim 1, characterized in that, The formula for calculating the difference includes: D(i,j)=|S(i,j)-T(i,j)| (2) Wherein, S(i,j) is the gray value corresponding to the pixel at position coordinate (i,j) in the template image, T(i,j) is the gray value corresponding to the pixel at position coordinate (i,j) in the transformed image, and D(i,j) is the gray value of the pixel at position coordinate (i,j) in the difference image.

Images