Joint detection method for surface defects of metal component

Abstract

The invention provides a metal component surface defect joint detection method, which relates to the field of metal detection, and comprises the following steps: S1, placing a workpiece; s2, eddy current detection; s3, transferring the workpiece; step S4, visual detection; and S5, judging the workpiece. The method integrates eddy current and mechanical vision to analyze and judge the surface defects of the thin-walled tubular or cylindrical metal component, and effectively solves the problems that a single machine vision detection method cannot accurately and quickly detect the surface defects of the metal component, detection blind areas are easy to appear, and the detection efficiency is high. And the problems of insensitivity, inaccurate quantification, easy signal interference and the like of single eddy current flaw detection on the surface defects of the metal components are solved, and efficient, accurate and comprehensive surface flaw detection of the metal components is realized.

Description

Technical Field

[0001] The invention relates to the technical field of metal detection, and in particular to a method for joint detection of surface defects of metal components. Background Art

[0002] Metal surface defects refer to discontinuous, irregular or non-compliant areas on the surface of metal materials during processing or use, including scratches, pits, pores, cracks, inclusions, etc. Metal surface defects can easily lead to reduced metal strength, weakened toughness, poor fatigue performance, accelerated corrosion and other problems. At present, the detection of metal surface defects mainly relies on manual inspection and machine vision inspection: Manual inspection, that is, manual inspection of metal surface defects with relevant inspection tools, which is difficult to detect, inefficient, low in detection accuracy, large in error, and greatly affected by human factors; Machine vision inspection, that is, converting the detected target into an image signal through an image capture device, transmitting it to a dedicated image processing system, converting it into a digital signal based on pixel distribution, brightness, color and other information, and identifying, analyzing and judging the defects on the metal surface through the digital signal and setting the threshold, so as to achieve the purpose of accurate detection. However, for some thin-walled tubular and cylindrical metal components, it is necessary not only to detect defects or damage on their surfaces, but also to detect defects or damage near their surfaces in order to effectively analyze the damage of the metal components and achieve the purpose of accurately detecting component defects. Although machine vision inspection can detect tiny surface defects with high sensitivity, it is unable to detect buried defects near the surface of thin-walled tubular and cylindrical metal components, resulting in insufficient comprehensive surface inspection of such metal components, prone to detection blind spots, and unable to accurately, efficiently and completely realize the analysis and judgment of defects of such metal components. Summary of the invention

[0003] In view of the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a joint detection method for surface defects of metal components. The method combines eddy current and machine vision to analyze and judge the surface defects of thin-walled tubular or cylindrical structure metal components, and effectively solves the problem that a single machine vision detection method cannot accurately and quickly detect the surface defects of such metal components, and is prone to detection blind spots.

[0004] The purpose of the present invention is achieved through the following technical solutions: A metal component surface defect joint detection method for metal components with thin-walled tubular and cylindrical structures, comprising: Step S1, workpiece placement: placing metal components on a conveying mechanism in sequence, and setting an eddy current detection module and a visual detection module on one side of the conveying mechanism, so as to facilitate eddy current detection and visual detection of the metal components in sequence; Step S2, eddy current detection: the metal component is moved to the eddy current detection module, and while the metal component rotates around its own axis, an outer diameter eddy current probe and an inner diameter eddy current probe are used to perform eddy current detection on the outer surface and inner surface of the metal component. During the detection process, the outer diameter eddy current probe and the inner diameter eddy current probe are moved simultaneously along the axis direction of the metal component; Step S3, workpiece transfer: according to the eddy current test results, the unqualified products are transferred to the unqualified area, and the qualified products are transferred to the visual inspection module; Step S4, visual inspection: while the qualified product of eddy current inspection rotates around its own axis, the outer diameter visual probe is started to perform visual inspection on the outer surface of the metal component. During the inspection, the outer diameter visual probe moves along the axis direction of the metal component; then, the metal component stops rotating, the inner diameter visual probe is started and moved along the axis direction of the metal component to perform visual inspection on the inner surface of the metal component; Step S5, workpiece judgment: first transfer the unqualified products detected by visual inspection to the unqualified area; then make a joint judgment on the qualified products detected by visual inspection, transfer the unqualified products detected by joint judgment to the unqualified area, and transfer the qualified products to the conveying mechanism for the next step of operation.

[0005] Based on further optimization of the above scheme, during the eddy current detection process of step S2, the distance between the end of the outer diameter eddy current probe and the outer surface of the metal component is 1 to 5 mm, and the distance between the end of the inner diameter eddy current probe and the inner surface of the metal component is 1 to 5 mm.

[0006] Based on the further optimization of the above scheme, the specific steps of eddy current detection in step S2 are: Step S21, scanning the outer surface and the inner surface of the metal component line by line along the axis direction by using the outer diameter eddy current probe and the inner diameter eddy current probe to obtain eddy current signal acquisition signals; Step S22, filtering processing: First, the time domain signal in the eddy current signal acquisition signal is processed to obtain the localized spectrum:

[0007] Where: represents a time series signal, w represents a real frequency variable; Then, the filtering interval is defined, and the harmonic components outside the filtering interval are reset to zero; Finally, the filtered frequency domain signal is converted into a time domain signal:

[0008] Step S23: extract the peak value of the signal after filtering and set the eddy current defect threshold F dIf the extracted signal peak is not less than the eddy current defect threshold, the metal component is judged to be unqualified; if the extracted signal peak is less than the eddy current defect threshold, the metal component is judged to be qualified, and the corresponding qualified product eddy current peak F is recorded.

[0009] Based on further optimization of the above solution, the filtering interval is 0 to 7 MHz.

[0010] Based on further optimization of the above scheme, during the visual inspection process of step S4, the distance between the end of the outer diameter visual probe and the corresponding outer surface of the metal component is 20 to 50 mm; the inner diameter visual probe is collinear with the central axis of the data component, and during the inspection process, the inner diameter visual probe moves from the outside to the inside of the metal component (that is, initially, the inner diameter visual probe is located outside the metal component).

[0011] Based on the further optimization of the above scheme, the specific steps of the visual inspection in step S4 are as follows: Step S41, scanning the outer surface and the inner surface of the metal component line by line along the axis direction by using the outer diameter visual probe and the inner diameter visual probe to obtain a visual signal acquisition image; Step S42: Image processing: First, the visual signal acquisition image is grayed out to obtain a grayscale image. The grayscale value of each pixel in the image for:

[0012] Where: R i 、G i 、B i Respectively represent i The color values ​​of the red channel, green channel, and blue channel of each pixel; Respectively represent the weight coefficients of the color values ​​corresponding to the red channel, green channel, and blue channel; After that, the grayscale image is denoised:

[0013] Where: (x,y) Indicates the coordinates of the corresponding pixel point; represents standard deviation; Then, the sobel operator is used to perform edge detection on the grayscale image after denoising, and the area surrounded by the upper edge, lower edge, left edge and right edge is divided to obtain the target image, and the rest of the image is removed as invalid background. Finally, set the binarization threshold T , perform binarization on the target image to obtain a binary image: like Gray(i)≥T , then the gray value of the pixel is 255. Gray(i)< T , then the gray value of the pixel is 0; Step S43: compare the obtained binary image with the preset standard defect binary image point by point. If the pixel matches, the pixel is marked as "1" and the coordinates of the corresponding pixel are obtained. (x i ,y i ) , otherwise it is recorded as "0"; Get the center point of all pixels marked as "1" (x 0 ,y 0 ) :

[0014] Center point (x 0 ,y 0 ) is the center of the circle, r A circular frame is drawn for the pixels marked as "1" with a radius of 0.001. If the number of pixels marked as "1" in the circular frame exceeds 2 / 3 of the area of ​​the corresponding circular frame, the area is judged as a defect, otherwise, it is not judged as a defect. Count the number and area of ​​all areas judged as defects, and preset the defect area threshold S d Defect count threshold N d If the total area of ​​the defective regions detected is not less than the defect area threshold and / or the number of defects is not less than the defect number threshold, the corresponding metal component is judged to be unqualified, and the total defect area S and the number of defects N are recorded; otherwise, the metal component is judged to be qualified.

[0015] Based on further optimization of the above scheme, the r It is 0.5 to 5 mm and is determined according to the scanning rate.

[0016] Based on the further optimization of the above scheme, the joint judgment in step S5 is specifically to obtain the joint judgment value of eddy current detection and visual detection P :

[0017] Where: a represents the eddy current peak equivalent, b Indicates the visual defect area equivalent, c Indicates the equivalent of the number of visual defects; Preset joint judgment threshold P dIf the joint judgment value is not less than the joint judgment threshold, the corresponding metal component is judged to be unqualified, otherwise, it is judged to be qualified.

[0018] The following are the technical effects of the solution of the present invention: This application combines the methods of first performing eddy current testing and then performing visual inspection, and through the complementary advantages of eddy current testing and visual inspection, effectively avoids the problems of single eddy current testing being insensitive to surface defect detection of metal components, inaccurate quantitative detection, and being susceptible to signal interference, as well as the inability of single visual inspection to detect defects below the surface and near-surface defects on the surface of metal components. It thereby achieves zero-blind zone detection of surface and near-surface defects of metal components, avoids detection blind spots, and improves detection accuracy and efficiency. This application interconnects and integrates eddy current and visual inspection data to, first, avoid the errors of a single detection method and improve the reliability and confidence of detection data; second, increase the detection range and avoid detection blind spots; and third, improve detection efficiency and avoid time-consuming and labor-intensive problems caused by multiple inspections, thereby effectively saving labor productivity and reducing production intensity. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flowchart of eddy current and visual joint detection in an embodiment of the present invention.

[0020] Figure 2 Schematic diagram of the structure of the joint detection system in Example 3 of the present invention.

[0021] Among them, 10. Conveying mechanism; 21. First rotating seat; 22. Servo drive mechanism for outer diameter eddy current probe; 23. Servo drive mechanism for inner diameter eddy current probe; 24. Outer diameter eddy current probe; 25. Inner diameter eddy current probe; 31. Second rotating seat; 32. Servo drive mechanism for outer diameter visual probe; 33. Servo drive mechanism for inner diameter visual probe; 34. Outer diameter visual probe; 35. Inner diameter visual probe; 41. Column; 42. First steering knuckle; 43. Second steering knuckle; 44. Gripper lifter; 45. Gripper; 50. Unqualified frame. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0023] Embodiment 1: A metal component surface defect joint detection method for metal components with thin-walled tubular and cylindrical structures, comprising: Step S1, workpiece placement: placing metal components on a conveying mechanism in sequence, and setting an eddy current detection module and a visual detection module on one side of the conveying mechanism, so as to facilitate eddy current detection and visual detection of the metal components in sequence; Step S2, eddy current detection: the metal component is moved to the eddy current detection module, and in the process of the metal component rotating around its own axis, an outer diameter eddy current probe and an inner diameter eddy current probe are used to perform eddy current detection on the outer surface and inner surface of the metal component. During the detection process, the distance between the end of the outer diameter eddy current probe and the outer surface of the metal component is 1 to 5 mm (preferably 3 mm), and the distance between the end of the inner diameter eddy current probe and the inner surface of the metal component is 1 to 5 mm (preferably 3 mm), and the outer diameter eddy current probe and the inner diameter eddy current probe are moved simultaneously along the axial direction of the metal component (the speed at which the outer diameter eddy current probe and the inner diameter eddy current probe move along the axial direction of the metal component matches the speed at which the metal component rotates around its own axis, so as to ensure that the eddy current flaw detection process covers a wide area); The specific steps of eddy current testing are: Step S21, scanning the outer surface and the inner surface of the metal component line by line along the axis direction by using the outer diameter eddy current probe and the inner diameter eddy current probe to obtain eddy current signal acquisition signals; Step S22, filtering processing: First, the time domain signal in the eddy current signal acquisition signal is processed to obtain the localized spectrum:

[0024] Where: represents a time series signal, w represents a real frequency variable; Then, the filter interval is defined, and the filter interval is 0 to 7MHz (the specific filter range is set according to the actual situation), and the harmonic components outside the filter interval are reset to zero; Finally, the filtered frequency domain signal is converted into a time domain signal:

[0025] Step S23: extract the peak value of the signal after filtering and set the eddy current defect threshold F d (Eddy current defect threshold F d Obtained from a large amount of empirical data in the laboratory), if the extracted signal peak is not less than the eddy current defect threshold, the metal component is judged to be unqualified; if the extracted signal peak is less than the eddy current defect threshold, the metal component is judged to be qualified, and the corresponding qualified product eddy current peak F is recorded.

[0026] Step S3, workpiece transfer: according to the eddy current test results, the unqualified products are transferred to the unqualified area, and the qualified products are transferred to the visual inspection module; Step S4, visual inspection: while the qualified product of eddy current inspection rotates around its own axis, the outer diameter visual probe is started to perform visual inspection on the outer surface of the metal component. During the inspection, the distance between the end of the outer diameter visual probe and the corresponding outer surface of the metal component is 20-50 mm (preferably 35 mm), and the outer diameter visual probe moves along the axis direction of the metal component (the movement speed of the outer diameter visual probe along the axis direction of the metal component matches the rotation speed of the metal component around itself, ensuring that the outer diameter visual probe can capture clear images without ghosting); thereafter, the metal component stops self-rotation, the inner diameter visual probe is started and moved along the axis direction of the metal component, and the inner surface of the metal component is visually inspected. The inner diameter visual probe is collinear with the central axis of the data component, and during the inspection, the inner diameter visual probe moves from the outside to the inside of the metal component (that is, initially, the inner diameter visual probe is located outside the metal component); The specific steps of visual inspection are: Step S41, scanning the outer surface and the inner surface of the metal component line by line along the axis direction by using the outer diameter visual probe and the inner diameter visual probe to obtain a visual signal acquisition image; Step S42: Image processing: First, the visual signal acquisition image is grayed out to obtain a grayscale image. The grayscale value of each pixel in the image for:

[0027] Where: R i 、G i 、B i Respectively represent i The color values ​​of the red channel, green channel, and blue channel of each pixel; Respectively represent the weight coefficients of the color values ​​corresponding to the red channel, the green channel, and the blue channel. In this embodiment, , proving that the green channel has the highest sensitivity, followed by the red channel, and the blue channel has the lowest sensitivity; After that, the grayscale image is denoised:

[0028] Where: (x,y) Indicates the coordinates of the corresponding pixel point; represents standard deviation; Then, the sobel operator is used to perform edge detection on the grayscale image after denoising, and the area surrounded by the upper edge, lower edge, left edge and right edge is divided to obtain the target image, and the rest of the image is removed as invalid background. Finally, set the binarization thresholdT (Binarization threshold T Obtained from a large amount of empirical data in the laboratory), the target image is binarized to obtain a binary image: like Gray(i)≥T , then the gray value of the pixel is 255. Gray(i)< T , then the gray value of the pixel is 0; Step S43: compare the obtained binary image with the preset standard defect binary image point by point. If the pixel matches, the pixel is marked as "1" and the coordinates of the corresponding pixel are obtained. (x i ,y i ) , otherwise it is recorded as "0"; Get the center point of all pixels marked as "1" (x 0 ,y 0 ) :

[0029] Center point (x 0 ,y 0 ) is the center of the circle, r The pixel points marked as "1" are framed with a circular shape. r The width is 0.5-5mm, which is determined according to the scanning rate. If the number of pixels marked as "1" in the circular frame exceeds 2 / 3 of the area of ​​the corresponding circular frame, the area is judged as a defect. Otherwise, it is not judged as a defect. Count the number and area of ​​all areas judged as defects, and preset the defect area threshold S d Defect count threshold N d (Defect area threshold S d , Defect quantity threshold N d Obtained from a large amount of empirical data in the laboratory), if the sum of the defect areas detected is not less than the defect area threshold and / or the number of defects is not less than the defect number threshold, the corresponding metal component is judged to be unqualified, and the sum of the defect areas S and the number of defects N are recorded; otherwise, the metal component is judged to be qualified.

[0030] Step S5, workpiece judgment: first transfer the unqualified products detected by visual inspection to the unqualified area; then make a joint judgment on the qualified products detected by visual inspection, specifically: obtain the joint judgment value of eddy current inspection and visual inspection P :

[0031] Where: a represents the eddy current peak equivalent, b Indicates the visual defect area equivalent, c Indicates the equivalent of the number of visual defects (a, b, c are obtained from a large amount of empirical data in the laboratory); Preset joint judgment threshold P d (Joint judgment threshold P d Obtained from a large amount of empirical data in the laboratory), if the joint judgment value is not less than the joint judgment threshold, the corresponding metal component is judged to be unqualified, otherwise, it is judged to be qualified.

[0032] Defective products are jointly judged and transferred to the defective area, and qualified products are transferred to the conveying mechanism for the next operation.

[0033] Embodiment 2: As another preferred embodiment of the present application, based on the solution in Example 1, the specific method for performing area statistics on all regions determined to be defects in step S43 is: First, the defect area is regarded as an approximate quadrilateral, and the distances from the pixel center point (point O) in four directions are obtained by Euclidean distance. (x 0 ,y 0 ) The farthest pixel coordinates: A ( x A ,y A )、B( x B ,y B )、C( x C ,y C )、B( x D ,y D ),

[0034] Then, the quadrilateral of the defect area is divided into four triangles: OAB, OBC, OCD, and OAD, and the areas of the four triangles are calculated by the vector method. M A 、M B 、M C 、M D ,For example:

[0035] After that, the total area of ​​the quadrilateral is obtained M :

[0036] And obtain the number of all pixels Z0 in the quadrilateral frame and the number of pixels corresponding to the defect area in the quadrilateral frame Z i , finally, the area of ​​the defect area is obtained M i : .

[0037] Embodiment 3: A combined detection system based on eddy current and machine vision, using the detection method in Example 1 or Example 2, referring to Figure 2 As shown: it includes a conveying mechanism 10, an eddy current detection module, a visual detection module, a transfer robot and a non-conforming frame 50, the eddy current detection module and the visual detection module are arranged on the same side of the conveying mechanism 10 and the transfer robot is arranged between the eddy current detection module and the visual detection module, and the non-conforming frame 50 is arranged on the side of the transfer robot away from the conveying mechanism; The conveying mechanism 10 adopts a conventional conveying production line (which can adopt a conveyor belt or other conveying methods) to convey the metal components to be inspected and the qualified metal components; The eddy current detection module includes a first rotating seat 21, an outer diameter eddy current probe servo drive mechanism 22, an inner diameter eddy current probe servo drive mechanism 23, an outer diameter eddy current probe 24 and an inner diameter eddy current probe 25, wherein the first rotating seat 21 is fixedly arranged on the ground corresponding to the transmission mechanism 10, and is used for supporting and self-rotating the metal component; the outer diameter eddy current probe servo drive mechanism 22 and the inner diameter eddy current probe servo drive mechanism 23 are arranged on the side of the first rotating seat 21 away from the transfer manipulator (such as Figure 2 As shown), the outer diameter eddy current probe servo drive mechanism 22 installs the outer diameter eddy current probe 24 through a mechanical arm, and the inner diameter eddy current probe servo drive mechanism 23 installs the inner diameter eddy current probe 25 through a mechanical arm (the mechanical arm can be a conventional mechanical arm, and the three-dimensional movement of the outer diameter eddy current probe 24 and the inner diameter eddy current probe 25 needs to be ensured), thereby respectively realizing the three-dimensional spatial movement of the outer diameter eddy current probe 24 and the inner diameter eddy current probe 25; The visual inspection module includes a second rotating seat 31, an outer diameter visual probe servo drive mechanism 32, an inner diameter visual probe servo drive mechanism 33, an outer diameter visual probe 34 and an inner diameter visual probe 35, wherein the second rotating seat 31 is fixedly arranged on the ground corresponding to the conveying mechanism 10, and is used for supporting and self-rotating the metal component; the outer diameter visual probe servo drive mechanism 32 and the inner diameter visual probe servo drive mechanism 33 are arranged on the side of the second rotating seat 31 away from the transfer robot (such as Figure 2As shown), the outer diameter visual probe servo drive mechanism 32 installs the outer diameter visual probe 34 through a mechanical arm, and the inner diameter visual probe servo drive mechanism 33 installs the inner diameter visual probe 35 through a mechanical arm (the mechanical arm can be a conventional mechanical arm, and the three-dimensional movement of the outer diameter visual probe 34 and the inner diameter visual probe 35 needs to be ensured), thereby respectively realizing the three-dimensional spatial movement of the outer diameter visual probe 34 and the inner diameter visual probe 35; The transfer robot comprises a column 41, a first steering knuckle 42, a second steering knuckle 43, a gripper lifter 44 and a gripper 45. The column 41 is fixedly arranged on the ground, the first steering knuckle 42 is rotatably arranged on the column 41, the second steering knuckle 43 is rotatably arranged at an end of the first steering knuckle 42 away from the column 41, the gripper lifter 44 is slidably arranged at an end of the second steering knuckle 43 away from the first steering knuckle 42, and the gripper 45 is fixedly arranged at the bottom end of the gripper lifter 44 (such as Figure 2 shown); The unqualified frame 50 is formed by welding cold-rolled sheet metal and is used to store the screened unqualified products.

Claims

1. A method for joint detection of surface defects of metal components, characterized by: include: Step S1, workpiece placement: placing metal components on a conveying mechanism in sequence, and an eddy current detection module and a visual detection module are respectively arranged on one side of the conveying mechanism; Step S2, eddy current detection: the metal component is moved to the eddy current detection module, and while the metal component rotates around its own axis, an outer diameter eddy current probe and an inner diameter eddy current probe are used to perform eddy current detection on the outer surface and inner surface of the metal component. During the detection process, the outer diameter eddy current probe and the inner diameter eddy current probe are moved simultaneously along the axis direction of the metal component; Step S3, workpiece transfer: according to the eddy current test results, the unqualified products are transferred to the unqualified area, and the qualified products are transferred to the visual inspection module; Step S4, visual inspection: while the qualified product of eddy current inspection rotates around its own axis, the outer diameter visual probe is started to perform visual inspection on the outer surface of the metal component. During the inspection, the outer diameter visual probe moves along the axis direction of the metal component; then, the metal component stops rotating, the inner diameter visual probe is started and moved along the axis direction of the metal component to perform visual inspection on the inner surface of the metal component; Step S5, workpiece judgment: first transfer the unqualified products detected by visual inspection to the unqualified area; then make a joint judgment on the qualified products detected by visual inspection, transfer the unqualified products detected by joint judgment to the unqualified area, and transfer the qualified products to the conveying mechanism for the next step of operation.

2. A metal component surface defect joint detection method according to claim 1, characterized in that: During the eddy current detection process of step S2, the distance between the end of the outer diameter eddy current probe and the outer surface of the metal component is 1 to 5 mm, and the distance between the end of the inner diameter eddy current probe and the inner surface of the metal component is 1 to 5 mm.

3. A metal component surface defect joint detection method according to claim 1 or 2, characterized in that: The specific steps of eddy current detection in step S2 are: Step S21, scanning the outer surface and the inner surface of the metal component line by line along the axis direction by using the outer diameter eddy current probe and the inner diameter eddy current probe to obtain eddy current signal acquisition signals; Step S22, filtering processing: First, the time domain signal in the eddy current signal acquisition signal is processed to obtain the localized spectrum: Where: represents a time series signal, w represents a real frequency variable; Then, the filtering interval is defined, and the harmonic components outside the filtering interval are reset to zero; Finally, the filtered frequency domain signal is converted into a time domain signal: Step S23: extract the peak value of the signal after filtering and set the eddy current defect threshold F d If the extracted signal peak is not less than the eddy current defect threshold, the metal component is judged to be unqualified; if the extracted signal peak is less than the eddy current defect threshold, the metal component is judged to be qualified, and the corresponding qualified product eddy current peak F is recorded.

4. A metal component surface defect joint detection method according to claim 3, characterized in that: During the visual inspection process of step S4, the distance between the end of the outer diameter visual probe and the corresponding outer surface of the metal component is 20 to 50 mm; the inner diameter visual probe is collinear with the central axis of the data component, and the inner diameter visual probe moves from the outside to the inside of the metal component during the inspection process.

5. A metal component surface defect joint detection method according to claim 4, characterized in that: The specific steps of the visual inspection in step S4 are: Step S41, scanning the outer surface and the inner surface of the metal component line by line along the axis direction by using the outer diameter visual probe and the inner diameter visual probe to obtain a visual signal acquisition image; Step S42: Image processing: First, the visual signal acquisition image is grayed out to obtain a grayscale image. The grayscale value of each pixel in the image for: Where: R i 、G i 、B i Respectively represent i The color values ​​of the red channel, green channel, and blue channel of each pixel; Respectively represent the weight coefficients of the color values ​​corresponding to the red channel, green channel, and blue channel; After that, the grayscale image is denoised: Where: (x,y) Indicates the coordinates of the corresponding pixel point; represents standard deviation; Then, the sobel operator is used to perform edge detection on the grayscale image after denoising, and the area surrounded by the upper edge, lower edge, left edge and right edge is divided to obtain the target image, and the rest of the image is removed as invalid background. Finally, set the binarization threshold T , perform binarization on the target image to obtain a binary image: like Gray(i)≥T , then the gray value of the pixel is 255. Gray(i)< T , then the gray value of the pixel is 0; Step S43: compare the obtained binary image with the preset standard defect binary image point by point. If the pixel matches, the pixel is marked as "1" and the coordinates of the corresponding pixel are obtained. (x i ,y i ) , otherwise it is recorded as "0"; Get the center point of all pixels marked as "1" (x 0 ,y 0 ) : Center point (x 0 ,y 0 ) is the center of the circle, r A circular frame is drawn for the pixels marked as "1" with a radius; if the number of pixels marked as "1" in the circular frame exceeds 2 / 3 of the area of ​​the corresponding circular frame, the area is judged as a defect, otherwise, it is not judged as a defect; Count the number and area of ​​all areas judged as defects, and preset the defect area threshold S d Defect count threshold N d If the total area of ​​the defective regions detected is not less than the defective area threshold and / or the number of defects is not less than the defective number threshold, the corresponding metal component is judged to be unqualified, and the total defective area S and the number of defects N are recorded; Otherwise, the metal component is judged to be qualified.

6. A metal component surface defect joint detection method according to claim 5, characterized in that: Said r 0.5~5mm.

7. A metal component surface defect joint detection method according to claim 5, characterized in that: The joint judgment in step S5 is specifically to obtain the joint judgment value of eddy current detection and visual detection. P : Where: a represents the eddy current peak equivalent, b Indicates the visual defect area equivalent, c Indicates the equivalent of the number of visual defects; Preset joint judgment threshold P d If the joint judgment value is not less than the joint judgment threshold, the corresponding metal component is judged to be unqualified, otherwise, it is judged to be qualified.

Images