Defect detection method of alloy anti-collision beam based on image processing
By constructing grayscale value curves and anomaly calculations, combined with Otsu method threshold segmentation, the efficiency and accuracy problems of traditional methods when detecting cracks of alloy anti-collision beams are solved, more accurate defect detection is achieved, and the safety of automobile production is improved Sex and efficiency.
Patent Information
- Application Number
- CN202411302787.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-19
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-09-19
AI Technical Summary
When detecting cracks in alloy anti-collision beams, traditional image processing methods reduce detection efficiency and accuracy due to the arc and the morphology of elongated cracks, and cannot accurately identify defect areas, causing product quality and safety hazards.
By constructing horizontal and vertical grayscale value curves, the detection value and anomaly of each pixel point are calculated, the image grayscale value is updated, and the crack area is accurately identified in combination with Otsu method threshold segmentation.
It improves detection efficiency and consistency, reduces the need for manual detection, reduces errors, and ensures the accuracy and safety of detection results.
Smart Images

Figure CN118823020B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of image processing, and more specifically, to an alloy anti-collision beam defect detection method based on image processing. Background Art
[0002] Aluminum alloy anti-collision beams are widely used in automobile manufacturing due to their light weight, high strength, good energy absorption performance and excellent corrosion resistance. Alloy anti-collision beam defect detection methods play an important role in automobile manufacturing and quality control; traditional manual detection methods will result in missed detections and false detections, and the detection results are highly subjective. At present, image processing technology and algorithms can be used to detect and analyze various defects that may exist in alloy anti-collision beams, and automatic detection can be performed, which greatly improves the efficiency and accuracy of detection.
[0003] The existing Chinese patent application document with publication number CN115861320A discloses an intelligent detection method for automobile parts processing information. The method obtains a grayscale image of the bearing surface after fluorescent magnetic powder spraying; obtains the suspected defect area in the grayscale image through a watershed algorithm; obtains the pixel point corresponding to the maximum grayscale value in each suspected defect area as the target pixel point, clusters the pixels in each suspected defect area, and takes the area of the cluster where the target pixel point is located as the target area; obtains the degree of fluorescent magnetic powder aggregation in each target area according to the area of the suspected defect area corresponding to the target area and the position distribution and gradient distribution of the pixels in the target area; obtains the grayscale difference between the target area and the grayscale image as the target difference; obtains the similarity between the target area and the background area according to the degree of fluorescent magnetic powder aggregation and the target difference; and determines the defect area in the grayscale image according to the similarity.
[0004] In the above scheme, fluorescent magnetic particle inspection may have specific requirements for environmental conditions (such as light, magnetic field, etc.), which may limit the inspection environment. At present, when defect detection of alloy anti-collision beams is performed through image processing, defect detection often uses Otsu threshold segmentation, but because the surface of the alloy anti-collision beam has curvature and the shape of the cracks is random and slender, the traditional Otsu threshold segmentation cannot accurately detect the existence of cracks, which reduces the inspection efficiency and accuracy, resulting in reduced product quality and safety hazards. Summary of the invention
[0005] In order to solve the problem that the surface of the alloy anti-collision beam has curvature and the cracks are random and slender, and the defect area cannot be accurately obtained during image detection, resulting in a decrease in detection accuracy, the present invention provides solutions in the following aspects.
[0006] The method for detecting defects of alloy anti-collision beams based on image processing comprises: obtaining an image of an alloy anti-collision beam to be detected, and performing preprocessing to obtain a grayscale image; constructing a transverse grayscale value curve and a longitudinal grayscale value curve based on the grayscale values of each row and column of pixels in the grayscale image, and calculating the detection value of each pixel on the transverse grayscale value curve and the longitudinal grayscale value curve; constructing a detection curve based on the detection value of each pixel, and using the difference in the detection values of the left and right adjacent pixels of each pixel in the detection curve to calculate the abnormality of each pixel; based on the difference between the abnormality of each pixel in the grayscale image and the abnormality of the pixels in the neighborhood, obtaining the relevant difference value corresponding to each pixel, and obtaining the defect degree of each pixel by performing correlation between the relevant difference value of each pixel in the alloy anti-collision beam image and the relevant difference value of the pixels at the symmetrical position, and updating the pixel of the alloy anti-collision beam image based on the defect degree of the pixel, and performing image segmentation on the updated alloy anti-collision beam image to obtain a crack area; wherein the defect degree satisfies the following relationship: ; In the formula, Indicates The defect degree of Indicates The pixel point corresponds to the Among the neighboring pixels of a pixel The relevant difference value of pixels, Indicates The pixel point corresponds to the Among the neighboring pixels of a pixel The relevant difference value of the symmetrical point of the pixel in the grayscale image, represents the normalization function, Indicates the number of pixels in the window. Indicates the number of pixel neighborhoods within the window.
[0007] The effect is: by constructing horizontal and vertical grayscale value curves and detection curves, the characteristics of the defects are reflected in detail, the defect degree and related difference values of the pixel points are used to update the image grayscale value, and the image of the crack area is further enhanced, so as to better identify the defective area, reduce the need for manual detection, improve detection efficiency and consistency, and reduce errors caused by human factors.
[0008] Preferably, constructing the transverse gray value curve and the longitudinal gray value curve includes:
[0009] The horizontal gray value curve is constructed by arranging the gray values of pixels in each row of the gray image from left to right, wherein the horizontal axis is the pixel number and the vertical axis is the pixel gray value;
[0010] The vertical grayscale value curve is constructed by arranging the pixel grayscale values of each column in the grayscale image in order from top to bottom, wherein the horizontal axis is the pixel number and the vertical axis is the pixel grayscale value.
[0011] The effect is that the horizontal and vertical grayscale value curves can intuitively display the laws and characteristics of grayscale distribution in the image. Abnormal fluctuations or mutations in the curves may indicate the presence of defects, which helps to quickly identify and locate potential defective areas in the image.
[0012] Preferably, calculating the detection value of each pixel point on the horizontal gray value curve and the vertical gray value curve includes:
[0013] Take any pixel as the target pixel, calculate the mean of the grayscale difference between the left and right sides of the target pixel on the horizontal grayscale value curve, and calculate the mean of the grayscale difference between the upper and lower sides of the target pixel on the vertical grayscale value curve, multiply the absolute value of the sum of the two means by the grayscale value of the target pixel, and perform normalization to obtain the detection value of the target pixel.
[0014] The effect is: by calculating the mean of the grayscale difference of the target pixel on the horizontal and vertical grayscale value curves, the degree of grayscale change around the pixel can be quantified. Larger grayscale differences are usually associated with defects. By averaging the left and right and up and down grayscale differences, the impact of random noise on the detection results can be reduced and the stability of the detection can be improved.
[0015] Preferably, constructing a detection curve according to the detection value of each pixel point includes:
[0016] The detection curve has a horizontal axis as a pixel point sequence and a vertical axis as the detection value of the corresponding pixel point.
[0017] Preferably, the abnormality of each pixel point is obtained by calculating the difference between the detection values of the left and right adjacent pixels in the detection curve, including:
[0018] Take any pixel point as the target pixel point, construct a window of a set size with the target pixel point as the center, calculate the difference in detection values of the left and right symmetrical pixels of the target pixel point in the window, calculate the sum of products after the difference and the slope of the target pixel point are normalized, and take the product of the sum of products and the detection value of the target pixel point as the abnormality of the target pixel point.
[0019] The effect is that by analyzing the difference between a pixel and its adjacent pixels on the grayscale value curve, it is possible to reveal abnormal or defective areas in the image. By obtaining the degree of abnormality, it can assist in image segmentation, and areas with significant grayscale changes are considered defective areas, thereby obtaining better segmentation effects.
[0020] Preferably, according to the correlation difference value between the abnormality of each pixel point in the alloy anti-collision beam image and the abnormality of the pixel point at the symmetrical position, it includes:
[0021] Taking any pixel point as the target pixel point, constructing a window of a set size with the target pixel point as the center, calculating the absolute value of the difference between the abnormality of two adjacent pixel points in the neighborhood pixel points of other pixel points other than the target pixel point in the window, and obtaining the slope value of the line between two adjacent pixel points in the neighborhood pixel points of other pixel points other than the target pixel point in the window, and taking the product of the absolute value of the difference and the slope value as the relevant difference value between the neighborhood pixel points of other pixel points other than the target pixel point in the window.
[0022] The effect is that by analyzing the abnormality difference and slope value of pixel pairs in the neighborhood of the target pixel point, the local image features can be captured more carefully, thereby improving the detection sensitivity of subtle defects.
[0023] Preferably, the pixels of the alloy anti-collision beam image are updated according to the defect degree of the pixels, satisfying the following relationship:
[0024] ;
[0025] In the formula, Indicates The defect degree of each pixel, Indicates The original gray value of the pixel, Indicates The new grayscale value of the pixel.
[0026] Preferably, performing image segmentation on the updated alloy anti-collision beam image to obtain the crack area includes:
[0027] The updated alloy anti-collision beam image is threshold segmented using the Otsu threshold method. The threshold segmentation method is: the grayscale image of the alloy anti-collision beam image is binarized, and the pixels in the grayscale image are divided into two categories: cracks and non-cracks, so as to obtain the crack area.
[0028] The present invention has the following effects:
[0029] 1. The present invention analyzes the characteristics of the alloy anti-collision beam, obtains the defect degree of each pixel according to the characteristics of the alloy anti-collision beam, and reallocates the grayscale value of each pixel according to the defect degree, so as to update the pixel grayscale of the entire image, and then uses the Otsu threshold method to perform threshold segmentation to obtain accurate crack pixel points and accurately detect the crack position.
[0030] 2. The present invention reallocates the grayscale value of pixels by defect degree, assists the Otsu threshold segmentation method, obtains accurate detection results, improves detection efficiency, discovers problems in time, and thus improves automobile production efficiency and ensures the quality and safety of safety structural components. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:
[0032] Figure 1 It is a method flow chart of steps S1 to S4 in the method for detecting defects of alloy anti-collision beams based on image processing in an embodiment of the present invention. DETAILED DESCRIPTION
[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0034] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.
[0035] Reference Figure 1 The alloy anti-collision beam defect detection method based on image processing includes steps S1 to S4, which are as follows:
[0036] S1: Obtain an image of the alloy anti-collision beam to be detected, and perform preprocessing to obtain a grayscale image.
[0037] Firstly, a high-definition camera is used to collect the image of the alloy anti-collision beam to be tested, and the image of the alloy anti-collision beam is preprocessed. The preprocessing adopts a semantic segmentation method to distinguish the alloy anti-collision beam from the background in the alloy anti-collision beam image, and obtain an image of the alloy anti-collision beam area. At this time, the image of the alloy anti-collision beam area is the image after removing the background, and it is grayscaled to obtain a grayscale image of the alloy anti-collision beam area.
[0038] It is understandable that because the grayscale of the crack pixel is random and affected by the crack, the grayscale of the pixels adjacent to the crack pixel will also change. Under normal circumstances, the anti-collision beam has an arc with high grayscale in the middle and low grayscale on both sides, and the grayscale of the pixels of the anti-collision beam is similar, so the pixels are judged by comparing the differences between the pixels.
[0039] S2: construct a horizontal grayscale value curve and a vertical grayscale value curve based on the grayscale value of each row and column of the grayscale image, and calculate the detection value of each pixel on the horizontal grayscale value curve and the vertical grayscale value curve.
[0040] The horizontal gray value curve is to arrange the gray values of each row of the gray image in order from left to right to construct a horizontal gray change curve, where the horizontal axis is the pixel number and the vertical axis is the pixel gray value;
[0041] The vertical gray value curve is to arrange the pixel gray values of each column in the gray image in order from top to bottom to construct a vertical gray change curve, where the horizontal axis is the pixel number and the vertical axis is the pixel gray value.
[0042] Take any pixel as the target pixel, calculate the mean of the grayscale difference between the left and right sides of the target pixel on the horizontal grayscale value curve, and calculate the mean of the grayscale difference between the upper and lower sides of the target pixel on the vertical grayscale value curve, multiply the absolute value of the sum of the two means by the grayscale value of the target pixel, and perform normalization to obtain the detection value of the target pixel.
[0043] Specifically, the detection value of the target pixel satisfies the following relationship:
[0044] ;
[0045] In the formula, Indicates The detection value of each pixel, Indicates The average of the left and right grayscale differences of the pixel points on the horizontal grayscale value curve, No. The average of the left and right grayscale differences of the pixel points on the vertical grayscale value curve, Represents pixel The gray value of Represents the normalization function.
[0046] In other words, the grayscale difference between the horizontal grayscale value curve and the vertical grayscale value curve and the grayscale value of the corresponding pixel point reflects the relationship when the horizontal and vertical grayscale curves have an influence on each other. The grayscale of the pixels of the anti-collision beam is similar, and the anti-collision beam has an arc with high grayscale in the middle and low grayscale on both sides.
[0047] In addition, in another feasible implementation, the average value of the variance of the grayscale values of all pixels in the corresponding row of the target pixel and the average value of the sum of the variance of the grayscale values of all pixels in the corresponding column of the target pixel can be used as the detection value of the target pixel.
[0048] On this basis, because the cracks are random and the shape of the cracks is elongated, when cracks occur, there are large differences in the detection values of the pixels. Therefore, a window is established with each pixel as the center, and judgment is made based on the differences in the pixels within the window.
[0049] S3: construct a detection curve according to the detection value of each pixel point, and use the difference in the detection values of the adjacent pixels on the left and right of each pixel point in the detection curve to calculate the abnormality of each pixel point.
[0050] The detection curve has the horizontal axis as the pixel sequence and the vertical axis as the detection value of the corresponding pixel.
[0051] Take any pixel as the target pixel, construct a window of set size with the target pixel as the center, calculate the difference of the detection values of the left and right symmetrical pixels of the target pixel in the window, calculate the product accumulation sum after the difference and the slope of the target pixel are normalized, and take the product of the product accumulation sum and the target pixel detection value as the abnormality of the target pixel.
[0052] In a feasible implementation manner, a 9×9 window is constructed with the target pixel point as the center, and the window may be adjusted according to actual conditions.
[0053] Specifically, the abnormality of the target pixel satisfies the following relationship:
[0054] ;
[0055] In the formula, Indicates The abnormality of each pixel, Indicates The detection value of each pixel, Indicates the window The left neighbor of the pixel The detection value of each pixel, Indicates the window The pixel to the right of The detection value of each pixel, Indicates the first detection curve in the window The slope of the pixel, Indicates the number of adjacent pixels.
[0056] That is to say, by analyzing the abnormality of each pixel point, due to the high lateral approximation of the anti-collision beam in the window, the pixels on the left and right sides of the target pixel point are selected to reflect the degree of influence of the cracks in the alloy anti-collision beam image. The larger the value, the stronger the influence of the cracks. By analyzing the slope of each pixel point in the window, the approximation of each pixel point in the window is reflected. It reflects the correlation of the symmetrical positions of pixels in the window. Under normal circumstances, each pixel is extremely similar, the detection value is relatively stable, and it shows a strong correlation. However, when an abnormal situation occurs, the detection value changes are more chaotic. The value range is , The value range is , After normalization, the range is , further processing, to ensure The normalized value range is , so it is converted into , yes Perform normalization.
[0057] In addition, in another feasible implementation, the abnormality of each pixel can be obtained by using the LOF algorithm. The specific steps are as follows:
[0058] A 9×9 window is constructed with the target pixel as the center. For example, the pixel spacing and the absolute value of the detection value difference between the target pixel and any other pixel in the window are selected in the 9×9 window, wherein the pixel spacing is obtained by the Euclidean distance between the target pixel and other pixel in the window, and the pixel spacing is used as the horizontal axis and the absolute value of the detection value difference is used as the vertical axis to obtain multiple two-dimensional data points (pixel spacing, absolute value of the detection value difference) of the target pixel. The two-dimensional data points are distributed in a rectangular coordinate system, and the two-dimensional data points are detected for anomalies using the LOF (Local Outlier Factor) algorithm. Set ,Specifically, the detection range can be adjusted according to the situation, and the LOF value corresponding to each 2D data point can be obtained, and the average of all LOF values can be ,normalized and used as the abnormality of the target pixel.
[0059] S4: Based on the difference between the abnormality of each pixel in the grayscale image and the abnormality of the pixels in the neighborhood, the relevant difference value corresponding to each pixel is obtained, and the correlation between the relevant difference value of each pixel in the alloy anti-collision beam image and the relevant difference value of the pixel at the symmetrical position is calculated to obtain the defect degree of each pixel. The pixel of the alloy anti-collision beam image is updated according to the defect degree of the pixel, and the updated alloy anti-collision beam image is segmented to obtain the crack area.
[0060] Taking any pixel as the target pixel, construct a window of a set size with the target pixel as the center, calculate the absolute value of the difference between the abnormality of two adjacent pixels in the neighborhood pixels of other pixels in the window other than the target pixel, and obtain the slope value of the line between two adjacent pixels in the neighborhood pixels of other pixels in the window other than the target pixel, and take the product of the absolute value of the difference and the slope value as the relevant difference value between the neighborhood pixels of other pixels in the window other than the target pixel.
[0061] Consistent with the window size set in the above step S3, a 9×9 window is constructed with the target pixel point as the center. Specifically, it can be adjusted according to actual conditions.
[0062] Specifically, the relevant difference value satisfies the following relationship:
[0063] ;
[0064] In the formula, Indicates The pixel point corresponds to the Among the neighboring pixels of a pixel The relevant difference value of pixels, Indicates The pixel point corresponds to the Among the neighboring pixels of a pixel Pixels and The slope of the line connecting the pixels, Indicates The pixel point corresponds to the Among the neighboring pixels of a pixel The abnormality of each pixel, Indicates The pixel point corresponds to the Among the neighboring pixels of a pixel The abnormality of each pixel.
[0065] That is, by calculating the correlation difference between two adjacent pixels in the neighborhood of other pixels other than the target pixel , the larger the correlation difference value is, the greater the abnormality in the corresponding window is.
[0066] Specifically, the defectivity satisfies the following relationship:
[0067] ;
[0068] In the formula, Indicates The defect degree of each pixel, Indicates The pixel point corresponds to the Among the neighboring pixels of a pixel The relevant difference value of pixels, Indicates The pixel point corresponds to the Among the neighboring pixels of a pixel The relevant difference value of the symmetrical point of the pixel in the grayscale image, represents the normalization function, Indicates the number of pixels in the window. Indicates the number of pixel neighborhoods within the window.
[0069] That is to say, the larger the defect value, the greater the possibility that the pixel is a crack pixel. , reaction The pixel window The larger the value, the worse the similarity is, and the greater the possibility of abnormality is.
[0070] On this basis, the grayscale value of each pixel is redistributed according to the defect degree. Since the grayscale value of the crack pixel is low, in order to increase the contrast between the crack and the normal position, the grayscale value of the crack pixel is reduced.
[0071] Specifically, the updated pixel points satisfy the following relationship:
[0072] ;
[0073] In the formula, Indicates The defect degree of Indicates The original gray value of the pixel, Indicates The new grayscale value of the pixel.
[0074] The updated alloy anti-collision beam image is threshold segmented using the Otsu threshold method. The threshold segmentation method is: the grayscale image of the alloy anti-collision beam image is binarized, and the pixels in the grayscale image are divided into two categories: cracks and non-cracks, so as to obtain the crack area.
[0075] Further explanation: Get the precise pixel points of the crack area and obtain the crack location. This can accurately evaluate the quality of the beam, improve the detection efficiency, accuracy, and timely detect problems, help manufacturers and quality control departments improve production efficiency, and ensure the quality and safety of safety structural components.
[0076] This is the end of the present invention.
[0077] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.
[0078] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.
Claims
1. The method for detecting defects of alloy anti-collision beams based on image processing is characterized in that: include: Acquire the image of the alloy anti-collision beam to be tested, and perform preprocessing to obtain a grayscale image; Using the grayscale values of each row and column of pixels in the grayscale image, construct a horizontal grayscale value curve and a vertical grayscale value curve, and calculate the detection value of each pixel on the horizontal grayscale value curve and the vertical grayscale value curve, including: taking any pixel as a target pixel, calculating the average of the grayscale differences on the left and right sides of the target pixel on the horizontal grayscale value curve, and calculating the average of the grayscale differences on the upper and lower sides of the target pixel on the vertical grayscale value curve, multiplying the absolute value of the sum of the two averages by the grayscale value of the target pixel, and performing normalization processing to obtain the detection value of the target pixel; A detection curve is constructed according to the detection value of each pixel point, and the difference in the detection values of the adjacent pixel points on the left and right of each pixel point in the detection curve is used to calculate the abnormality of each pixel point, including: taking any pixel point as the target pixel point, constructing a window of a set size with the target pixel point as the center, calculating the difference in the detection values of the left and right symmetrical pixel points of the target pixel point in the window, calculating the product accumulation sum after the difference is normalized with the slope of the target pixel point, and taking the product of the product accumulation sum and the detection value of the target pixel point as the abnormality of the target pixel point; Based on the difference between the abnormality of each pixel point in the grayscale image and the abnormality of the pixels in the neighborhood, the relevant difference value corresponding to each pixel point is obtained, including: taking any pixel point as the target pixel point, constructing a window of a set size with the target pixel point as the center, calculating the absolute value of the difference between the abnormality of two adjacent pixels in the neighborhood pixels of other pixels other than the target pixel point in the window, and obtaining the slope value of the line between two adjacent pixels in the neighborhood pixels of other pixels other than the target pixel point in the window and normalizing it, and taking the product of the absolute value of the difference and the normalized slope value as the relevant difference value between the neighborhood pixels of other pixels other than the target pixel point in the window; the correlation between the relevant difference value of each pixel point in the alloy anti-collision beam image and the relevant difference value of the pixel point at the symmetrical position is obtained to obtain the defect degree of each pixel point; updating the pixel point of the alloy anti-collision beam image according to the defect degree of the pixel point to satisfy the following relationship: ; In the formula, Indicates The defect degree of each pixel, Indicates The original gray value of the pixel, Indicates The new gray value of each pixel point is obtained; the updated alloy anti-collision beam image is segmented to obtain the crack area; Wherein, the defect degree satisfies the following relationship: ; In the formula, Indicates The pixel point corresponds to the Among the neighboring pixels of a pixel The relevant difference value of pixels, Indicates The pixel point corresponds to the Among the neighboring pixels of a pixel The relevant difference value of the symmetrical point of the pixel in the grayscale image, represents the normalization function, Indicates the number of pixels in the window. Indicates the number of pixel neighborhoods within the window.
2. The method for detecting defects of alloy anti-collision beams based on image processing according to claim 1, characterized in that: Constructing a horizontal gray value curve and a vertical gray value curve, including: The horizontal gray value curve is constructed by arranging the gray values of pixels in each row of the gray image from left to right, wherein the horizontal axis is the pixel number and the vertical axis is the pixel gray value; The vertical grayscale value curve is constructed by arranging the pixel grayscale values of each column in the grayscale image in order from top to bottom, wherein the horizontal axis is the pixel number and the vertical axis is the pixel grayscale value.
3. The method for detecting defects of alloy anti-collision beams based on image processing according to claim 1, characterized in that: Construct a detection curve based on the detection value of each pixel. include: The detection curve has a horizontal axis as a pixel point sequence and a vertical axis as the detection value of the corresponding pixel point.
4. The method for detecting defects of alloy anti-collision beams based on image processing according to claim 1, characterized in that: Perform image segmentation on the updated alloy anti-collision beam image to obtain the crack area, including: The updated alloy anti-collision beam image is threshold segmented using the Otsu threshold method. The threshold segmentation method is: the grayscale image of the alloy anti-collision beam image is binarized, and the pixels in the grayscale image are divided into two categories: cracks and non-cracks, so as to obtain the crack area.
Citation Information
Patent Citations
Intelligent detection method for automobile part machining information
CN115861320A
Visual inspection system for workpiece appearance defects based on image processing
CN118470016A
