Glass bead defect detection method based on optimized image analysis algorithm

Abstract

The invention relates to the technical field of image data processing, and provides a glass bead defect detection method based on an optimized image analysis algorithm, which comprises the following steps: acquiring multi-angle image data of a glass bead; based on the multi-angle image data of the glass beads, performing first defect detection on the glass beads: if the glass beads are detected to have defects, marking the glass beads as defective glass beads; otherwise, the multi-angle image data is substituted into a defect detection module, so that two-dimensional detection or three-dimensional detection is selectively performed on the glass beads according to the multi-angle image data to obtain detection results of the glass beads, and the detection results comprise whether the glass beads have defects or not and defect types of the glass beads when the glass beads have the defects; and marking the glass beads as defective glass beads based on the detection results of the glass beads. The image acquisition comprehensiveness of the glass beads is improved, the defect condition of the glass beads can be accurately detected, and the detection accuracy and the detection efficiency of the glass beads are further improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image data processing, in particular to the technical field of glass bead defect detection, and specifically provides a glass bead defect detection method based on an optimized image analysis algorithm. BACKGROUND

[0002] Glass beads are ornaments accessories with silica as the main material, which are widely used in clothing, handicrafts and industrial fields. However, in the glass bead production process, defects in glass beads are prone to occur;

[0003] In the prior art, the glass bead defect detection technology depends on image accuracy. However, in the glass bead production process, environmental vibration and detection device shaking have an impact on measurement accuracy. Therefore, the glass bead defect detection efficiency is very low, and the glass bead defect detection technology has limitations.

[0004] Correspondingly, there is a need for a new glass bead defect detection scheme based on an optimized image analysis algorithm to solve the above problems. SUMMARY

[0005] In order to overcome the above defects, the present application is proposed to provide a glass bead defect detection method based on an optimized image analysis algorithm to solve or at least partially solve the technical problems in the prior art that environmental vibration and detection device shaking have an impact on measurement accuracy, resulting in very low glass bead defect detection efficiency and leading to limitations of glass bead defect detection technology.

[0006] In a first aspect, the present application provides a glass bead defect detection method based on an optimized image analysis algorithm, comprising the following steps:

[0007] Obtaining multi-angle image data of a glass bead;

[0008] Based on the multi-angle image data of the glass bead, performing a first defect detection on the glass bead:

[0009] If it is detected that there is a defect, the glass bead is marked as a defective glass bead;

[0010] Otherwise, the multi-angle image data is substituted into a defect detection module, so that according to the multi-angle image data, a two-dimensional detection or a three-dimensional detection is selectively performed on the glass bead, and a detection result of the glass bead is obtained, wherein the detection result includes whether the glass bead has a defect, and when the glass bead has a defect, the defect type of the glass bead;

[0011] Based on the test results of the glass beads, the glass beads are marked as defective glass beads.

[0012] In one technical solution of the glass bead defect detection method based on the above-mentioned optimized image analysis algorithm, the step of substituting the multi-angle image data into the defect detection module, so that the glass bead is selectively subjected to two-dimensional or three-dimensional detection based on the multi-angle image data, to obtain a detection result of whether the glass bead has defects, includes:

[0013] Based on the multi-angle image data, detect whether there are interference items in the multi-angle image data of the glass bead:

[0014] If it is determined that it exists, then the "acquire multi-angle image data of glass bead" is executed again, and the number of times the multi-angle image data of glass bead is acquired, as well as the subsequent steps, are recorded;

[0015] If it is determined that the glass bead does not exist, or if the number of times the multi-angle image of the glass bead is acquired is 2, then the glass bead is fused based on all the multi-angle image data of the glass bead to obtain the fused image of the glass bead.

[0016] Based on the fused image of the glass beads, two-dimensional or three-dimensional defect detection is selectively performed on the glass beads.

[0017] In one technical solution of the glass bead defect detection method based on the above-mentioned optimized image analysis algorithm, the selective two-dimensional or three-dimensional defect detection of the glass bead based on the fused image of the glass bead includes:

[0018] Based on the fused image of the glass beads, a second defect detection is performed on the glass beads:

[0019] If a defect is detected, the defect type of the glass bead is determined, and a detection result indicating that the glass bead has a defect is generated, thereby achieving two-dimensional detection of the glass bead.

[0020] If no defects are detected, a three-dimensional surface model of the glass beads is generated based on the fused image of the glass beads.

[0021] Based on the three-dimensional surface model of the glass bead, determine whether the glass bead has defects:

[0022] If the defect is determined to exist, the defect type of the glass bead is determined, and the detection result of the glass bead is generated to achieve three-dimensional detection of the glass bead.

[0023] Otherwise, a test result indicating that the glass bead has no defects is generated.

[0024] In one technical solution of the glass bead defect detection method based on the above-mentioned optimized image analysis algorithm, the step of detecting whether there are interference items in the multi-angle image data of the glass bead based on the multi-angle image data includes:

[0025] The degree of image interference can be determined using the following formula:

[0026]

[0027] in Representative image The degree of interference, Represents the image spatial domain. Represents the image at location The Laplace operator, The mean of the Laplace operator, The weight coefficients representing feature point matching. as well as These represent the first and second images in the reference and test images, respectively. One feature point, The weighting coefficients representing the frequency domain influence of vibration. Represents frequency variables. Represents the Fourier transform of the image. Represents the parameters of the vibration transfer function;

[0028] when When the multi-angle image data of the glass bead is detected, it is determined that there are interference items. This represents the preset threshold for the degree of interference, with a value of [value missing]. ;

[0029] Otherwise, it is determined that there are no interference items in the multi-angle image data of the glass bead being detected.

[0030] In one technical solution of the glass bead defect detection method based on the above-mentioned optimized image analysis algorithm, the step of fusing all multi-angle image data of the glass bead to obtain a fused image of the glass bead includes:

[0031] The fusion image of the glass beads is obtained using the following formula:

[0032]

[0033] in, Represents the merged image. The first representing glass beads The image of the frame, The weighting coefficients representing the smoothing term. The value ranges from [0.01, 1.0]. The weight coefficients representing the Huber loss function. The value ranges from [0.1, 1.0]. This represents the Huber loss function value.

[0034] In one technical solution of the glass bead defect detection method based on the above-mentioned optimized image analysis algorithm, determining whether the glass bead has defects based on the three-dimensional surface model of the glass bead includes:

[0035] The following formula is used to monitor whether the glass beads have defects:

[0036]

[0037] in, This represents the energy functional for defect detection. Represents the defect probability field. A three-dimensional surface model representing a glass bead. Represents the gradient term The weighting coefficients, Representing Laplace's term The weighting coefficients, The weighting coefficients represent the boundary terms. Represents boundary conditions. Represents the three-dimensional volume region inside the bubble. Represents the surface boundary of the bubble;

[0038] Defect detection is performed on the glass beads based on the defect detection energy functional.

[0039] In one technical solution of the glass bead defect detection method based on the above-mentioned optimized image analysis algorithm, the step of determining the defect type of the glass bead and generating the detection result of the glass bead to achieve three-dimensional detection of the glass bead includes:

[0040] The defect type of the glass bead is determined by the following formula:

[0041]

[0042] in, Represents the defect type of the glass bead. The feature vector representing the defect to be tested. Representative characteristics template feature vector, Representing the Each feature similarity parameter as well as These represent the classifier parameters.

[0043] In one technical solution of the glass bead defect detection method based on the above-mentioned optimized image analysis algorithm, the feature vector of the defect to be tested includes at least bubble vector, discoloration vector, protrusion and pit vector, scar and white mark and wear vector;

[0044] The bubble vector in the feature vector of the defect to be tested is obtained by the following formula:

[0045]

[0046] in, Represents the bubble vector. Represents the curvature of the surface inside the bubble. Represents the bubble density field. Represents the divergence of the normal vector. Represents the boundary term weight coefficients;

[0047] The heterochromatic vector in the feature vector of the defect to be tested is obtained by the following formula:

[0048]

[0049] in, Represents a heterochromatic vector. Represents the square of the color gradient magnitude. Represents the energy of the color spectrum. The Laplacian operator represents the color. as well as These represent the weighting coefficients.

[0050] In one technical solution of the glass bead defect detection method based on the above-mentioned optimized image analysis algorithm, the convex and concave vectors in the feature vector of the defect to be tested are obtained by the following formula:

[0051]

[0052] in, Represents the vectors of bulges and depressions. The curvature representing the protruding defect. The curvature representing a protruding defect is a positive value. The curvature representing a protruding defect is negative. The average value representing the background curvature;

[0053] The scar, white mark, and wear vector in the feature vector of the defect to be tested are obtained by the following formula:

[0054]

[0055] in, This represents the squared magnitude of the color gradient, used to reflect color abrupt changes in scars / white marks. This represents the second-order color gradient, used to reflect the gradual color change in the wear area. Represents the curvature of the defective region. Represents the curvature gradient;

[0056] Furthermore, through the matrix The combination of element values ​​determines the defect type of scars, white marks, and wear vectors:

[0057] 1) Scars: High color gradient, high curvature, high curvature gradient;

[0058] 2) White print: High color gradient, low curvature, low curvature gradient;

[0059] 3) Wear: Low color gradient, low curvature, low curvature gradient.

[0060] In one technical solution of the above-mentioned glass bead defect detection method based on optimized image analysis algorithm, the method further includes:

[0061] During the detection process of the defect detection module, a comprehensive score is given for the detection process of any number of glass beads randomly selected, to ensure the accuracy of the defect detection module and to optimize the defect detection module in a timely manner.

[0062] The comprehensive score for the glass bead testing process is obtained using the following formula:

[0063]

[0064] in, The index represents the evaluation index of fusion effect. This represents the index for evaluating the effectiveness of defect detection. , , , All represent weighting coefficients, satisfying , This represents the gradient field of the overall score, used to constrain the smoothness of the score.

[0065] The above-described technical solutions of the present invention have at least one or more of the following beneficial effects:

[0066] In implementing the technical solution of this invention, by acquiring multi-angle image data of the glass bead, the comprehensiveness of image acquisition of the glass bead is improved. Furthermore, the glass bead is subjected to a first defect detection using this multi-angle image data, achieving direct defect detection. If no defect is detected in the first defect detection, the multi-angle image data is substituted into the defect detection module, and the glass bead is selectively subjected to two-dimensional or three-dimensional detection to obtain the detection result. This enables multiple accurate detections of the glass bead's defects, thereby improving the detection accuracy and efficiency. It avoids the technical problem in existing glass bead defect detection technologies where environmental vibration and equipment shaking affect measurement accuracy, resulting in low defect detection efficiency and limitations. Attached Figure Description

[0067] The disclosure of this invention will become more readily understood with reference to the accompanying drawings. It will be readily understood by those skilled in the art that these drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. Furthermore, similar numbers in the drawings are used to denote similar components, wherein:

[0068] Figure 1 This is a schematic flowchart of the main steps of a glass bead defect detection method based on an optimized image analysis algorithm according to an embodiment of the present invention;

[0069] Figure 2 This is a flowchart illustrating the defect detection module of a glass bead defect detection method based on an optimized image analysis algorithm according to an embodiment of the present invention.

[0070] Figure 3 This is a schematic diagram of a glass bead defect detection device based on an optimized image analysis algorithm, according to an embodiment of the present invention.

[0071] List of reference numerals in the attached diagram:

[0072] 1: Glass bead defect detection device based on optimized image analysis algorithm; 11: Feeding module; 12: Defective product unloading bin. Detailed Implementation

[0073] Some embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0074] In the description of this invention, "module" and "processor" can include hardware, software, or a combination of both. A module can include hardware circuitry, various suitable sensors, communication ports, memory, and may also include software components, such as program code, or a combination of software and hardware. A processor can be a central processing unit, microprocessor, image processor, digital signal processor, or any other suitable processor. The processor has data and / or signal processing capabilities. The processor can be implemented in software, in hardware, or a combination of both. Non-transitory computer-readable storage media includes any suitable medium capable of storing program code, such as magnetic disks, hard disks, optical disks, flash memory, read-only memory, random access memory, etc. The term "A and / or B" means all possible combinations of A and B, such as only A, only B, or A and B. The terms "at least one A or B" or "at least one of A and B" have a similar meaning to "A and / or B" and can include only A, only B, or A and B. The singular terms "a" or "this" can also include plural forms.

[0075] See appendix Figure 1 , Figure 1 This is a schematic flowchart illustrating the main steps of a glass bead defect detection method based on an optimized image analysis algorithm according to an embodiment of the present invention. Figure 1 As shown, the glass bead defect detection method based on optimized image analysis algorithm in this embodiment of the invention mainly includes the following steps S101-S104.

[0076] Step S101: Acquire multi-angle image data of the glass bead;

[0077] Specifically, the multi-angle image data of the glass bead is acquired by using high frame rate cameras to capture images of the glass bead, wherein the number of cameras is at least 8.

[0078] Step S102: Based on the multi-angle image data of the glass bead, perform the first defect detection on the glass bead:

[0079] Step S103: If a defect is detected, the glass bead is marked as a defective glass bead;

[0080] Otherwise, the multi-angle image data is substituted into the defect detection module, so that the glass bead is selectively inspected in two dimensions or three dimensions based on the multi-angle image data, and the detection result of the glass bead is obtained. The detection result includes whether the glass bead has a defect, and when the glass bead has a defect, the type of defect of the glass bead.

[0081] like Figure 2As shown, specifically, the step of substituting the multi-angle image data into the defect detection module, so that the glass bead is selectively subjected to two-dimensional or three-dimensional detection based on the multi-angle image data, to obtain a detection result indicating whether the glass bead has defects includes:

[0082] Based on the multi-angle image data, detect whether there are interference items in the multi-angle image data of the glass bead:

[0083] If it is determined that it exists, then the "acquire multi-angle image data of glass bead" is executed again, and the number of times the multi-angle image data of glass bead is acquired, as well as the subsequent steps, are recorded;

[0084] If it is determined that the glass bead does not exist, or if the number of times the multi-angle image of the glass bead is acquired is 2, then the glass bead is fused based on all the multi-angle image data of the glass bead to obtain the fused image of the glass bead.

[0085] Based on the fused image of the glass beads, two-dimensional or three-dimensional defect detection is selectively performed on the glass beads.

[0086] In the above embodiments, by detecting whether there are interference items in the multi-angle image data of the glass bead, the multi-angle image data of the glass bead is selectively reacquired. This enables the reacquisition of image data when environmental vibrations and shaking of the detection equipment affect the measurement accuracy, thereby improving the accuracy of the glass bead's image data. When multi-angle image data has been acquired multiple times, or when there are no interference items, the image data of the glass bead is fused to generate a fused image. Based on this fused image, the glass bead can be selectively subjected to two-dimensional or three-dimensional defect detection, thereby improving the accuracy of defect detection of the glass bead.

[0087] Specifically, the fused image based on glass beads, selectively performing two-dimensional or three-dimensional defect detection on the glass beads, includes:

[0088] Based on the fused image of the glass beads, a second defect detection is performed on the glass beads:

[0089] If a defect is detected, the defect type of the glass bead is determined, and a detection result indicating that the glass bead has a defect is generated, thereby achieving two-dimensional detection of the glass bead.

[0090] If no defects are detected, a three-dimensional surface model of the glass beads is generated based on the fused image of the glass beads.

[0091] Based on the three-dimensional surface model of the glass bead, determine whether the glass bead has defects:

[0092] If the defect is determined to exist, the defect type of the glass bead is determined, and the detection result of the glass bead is generated to achieve three-dimensional detection of the glass bead.

[0093] Otherwise, a test result indicating that the glass bead has no defects is generated.

[0094] In the above embodiments, a second defect detection is performed on the fused image of the glass bead, realizing two-dimensional detection of the glass bead. When no defect is detected, a three-dimensional surface model of the glass bead is generated, and a third defect detection is performed on the glass bead, realizing three-dimensional detection of the glass bead. By performing three defect detections on the glass bead, the accuracy of defect detection is ensured, while the efficiency of defect detection is improved. Each time a defect is detected, the defect type of the glass bead is determined, so as to facilitate the review of the glass bead detection process, thereby improving the self-inspection and optimization capabilities of the defect detection module.

[0095] Specifically, detecting whether there are interference items in the multi-angle image data of the glass bead based on the multi-angle image data includes:

[0096] The degree of image interference can be determined using the following formula:

[0097]

[0098] in, Representative image The degree of interference, Represents the image spatial domain. Represents the image at location The Laplacian operator is used to reflect edge information. The mean of the Laplace operator, The weight coefficients representing feature point matching. as well as These represent the first and second images in the reference and test images, respectively. One feature point, The weighting coefficients representing the frequency domain influence of vibration. Represents frequency variables. Represents the Fourier transform of the image. Represents the vibration transfer function parameters, where, ,in Represents the system's critical frequency;

[0099] when When the multi-angle image data of the glass bead is detected, it is determined that there are interference items. This represents the preset threshold for the degree of interference, with a value of [value missing]. ;

[0100] Otherwise, it is determined that there are no interference items in the multi-angle image data of the glass bead being detected.

[0101] Specifically, the process of fusing all multi-angle image data of the glass bead to obtain a fused image of the glass bead includes:

[0102] The fusion image of the glass beads is obtained using the following formula:

[0103]

[0104] in, Represents the merged image. The first representing glass beads The image of the frame, The weighting coefficients representing the smoothing term. The value ranges from [0.01, 1.0]. The weight coefficients representing the Huber loss function. The value ranges from [0.1, 1.0]. This represents the Huber loss function value.

[0105] Specifically, the Huber loss function value is obtained using the following formula:

[0106]

[0107] in, Represents the threshold, with a value of [0.01]. 1.0 ].

[0108] Specifically, determining whether the glass bead has defects based on its three-dimensional surface model includes:

[0109] The following formula is used to monitor whether the glass beads have defects:

[0110]

[0111] in, This represents the energy functional for defect detection. Represents the defect probability field. A three-dimensional surface model representing a glass bead. Represents the gradient term The weighting coefficients, Representing Laplace's term The weighting coefficients, The weighting coefficients represent the boundary terms. Represents boundary conditions. Representing a three-dimensional volume region inside the bubble, used to quantify curvature-density coupling anomalies within the volume. Represents the surface boundary of a bubble, used to quantify discontinuities at the boundary;

[0112] Defect detection is performed on the glass beads based on the defect detection energy functional.

[0113] Specifically, defect detection of the glass bead based on the defect detection energy functional includes:

[0114] Determine the energy functional value of a defect-free sample ;

[0115] Energy threshold It can be obtained through the following formula:

[0116]

[0117] in, as well as These represent the mean and standard deviation, respectively. Represents the empirical coefficient, which is taken here. ;

[0118] When the glass beads If so, the glass bead is determined to have a defect.

[0119] Specifically, in some embodiments, if any region of the glass bead Significantly larger than other regions (i.e.) and If the difference is large, then the area is determined to have a defect.

[0120] If any region of the glass bead is in the smoothness term ( If a region cannot be fitted under certain constraints (i.e., significant gradient or curvature change must be preserved), then the region is considered to have a defect.

[0121] If any region boundary of the glass bead If the difference is significant (such as a protruding or recessed boundary), then the area is determined to have a defect.

[0122] Specifically, determining the defect type of the glass bead and generating the detection result of the glass bead to achieve three-dimensional detection of the glass bead includes:

[0123] The defect type of the glass bead is determined by the following formula:

[0124]

[0125] in, Represents the defect type of the glass bead. The feature vector representing the defect to be tested. Representative characteristics template feature vector, Representing the Each feature similarity parameter as well as These represent the classifier parameters;

[0126] Specifically, That is, the first The weights of the kernel function in the decision function are used to describe the th kernel function. The degree of contribution of each kernel function to the classification result, It can be set by those skilled in the art according to actual usage needs, or it can be obtained through optimization algorithms; This refers to the bias term of the decision function, which is used to adjust the position of the classification boundary to ensure that the sample is correctly classified. It can be set by those skilled in the art based on actual usage needs, or it can be obtained through optimization algorithms.

[0127] Specifically, the feature vector of the defect to be tested includes at least bubble vector, discoloration vector, protrusion and pit vector, scar and white mark and wear vector;

[0128] The bubble vector in the feature vector of the defect to be tested is obtained by the following formula:

[0129]

[0130] in, Represents the bubble vector. Represents the curvature of the surface inside the bubble. Represents the bubble density field. Represents the divergence of the normal vector. Represents the boundary term weight coefficients;

[0131] The heterochromatic vector in the feature vector of the defect to be tested is obtained by the following formula:

[0132]

[0133] in, Represents a heterochromatic vector. Represents the square of the color gradient magnitude. Represents the energy of the color spectrum. The Laplacian operator represents the color. as well as These represent the weighting coefficients.

[0134] Specifically, the convex and concave vectors in the feature vector of the defect to be tested are obtained by the following formula:

[0135]

[0136] in, Represents the vectors of bulges and depressions. The curvature representing the protruding defect. The curvature representing a protruding defect is a positive value. The curvature representing a protruding defect is negative. The average value representing the background curvature;

[0137] The scar, white mark, and wear vector in the feature vector of the defect to be tested are obtained by the following formula:

[0138]

[0139] in, This represents the squared magnitude of the color gradient, used to reflect color abrupt changes in scars / white marks. This represents the second-order color gradient, used to reflect the gradual color change in the wear area. Represents the curvature of the defective region. Represents the curvature gradient;

[0140] Furthermore, through the matrix The combination of element values ​​determines the defect type of scars, white marks, and wear vectors:

[0141] 1) Scars: High color gradient, high curvature, high curvature gradient;

[0142] 2) White print: High color gradient, low curvature, low curvature gradient;

[0143] 3) Wear: Low color gradient, low curvature, low curvature gradient.

[0144] Specifically, the comparison objects for the above high and low are the background color, curvature, and curvature gradient of the glass beads.

[0145] Specifically, the average value of the background curvature It can be obtained through the following formula:

[0146]

[0147] in, The surface curvature representing the defect-free region. This represents the overall statistical background area.

[0148] Specifically, the defect types of the glass beads include at least bubbles, discoloration, protrusions, pits, scars, white marks, and wear.

[0149] Specifically, bubble vector Used to determine the severity of bubble defects. The higher the bubble, the larger it is, the more distorted its shape, or the clearer its boundaries.

[0150] Different color vectors are used to determine different color regions and the severity of different colors.

[0151] Specifically, the method further includes:

[0152] During the detection process of the defect detection module, a comprehensive score is given for the detection process of any number of glass beads randomly selected, to ensure the accuracy of the defect detection module and to optimize the defect detection module in a timely manner.

[0153] The comprehensive score for the glass bead testing process is obtained using the following formula:

[0154]

[0155] in, The index represents the evaluation index of fusion effect. This represents the index for evaluating the effectiveness of defect detection. , , , All represent weighting coefficients, satisfying , This represents the gradient field of the overall score, used to constrain the smoothness of the score.

[0156] Specifically, the fusion effect evaluation index It can be obtained through the following formula:

[0157]

[0158] in, This represents the structural similarity index between the fused image and the real image, with values ​​ranging from [0,1]. Represents the merged image. The image represents the non-interference item. The image represents the presence of interfering terms. This represents the variance of the Laplacian operator, used to reflect the edge sharpness of the image. , , , These represent the weighting coefficients, respectively, satisfying... .

[0159] Specifically, the defect detection effectiveness evaluation index It can be obtained through the following formula:

[0160]

[0161] in, The F1 score represents the defect detection result. The heterochromatic vector representing the detected defect. Represents the actual defect color vector. Represents the three-dimensional Hausdorff distance. Represents a collection of point clouds. Represents the maximum pixel spacing. Represents the curvature of the surface. Represents the defective area. , , , These represent the weighting coefficients, respectively, satisfying... .

[0162] Step S104: Based on the test results of the glass beads, mark the glass beads as defective glass beads.

[0163] Specifically, after marking the glass beads as defective glass beads based on the test results, the method further includes:

[0164] Each defect-free glass bead is placed in the position of the good material;

[0165] Each defective glass bead is placed in the location of the defective material.

[0166] Based on steps S101-S104 above, by acquiring multi-angle image data of the glass bead, the comprehensiveness of image acquisition of the glass bead is improved. Furthermore, the glass bead is subjected to a first defect detection using this multi-angle image data, realizing direct defect detection of the glass bead. If no defect is detected in the first defect detection, the multi-angle image data is substituted into the defect detection module, and the glass bead is selectively subjected to two-dimensional or three-dimensional detection to obtain the detection result of the glass bead. This enables multiple accurate detections of the glass bead's defect status, thereby improving the detection accuracy and efficiency of the glass bead. It avoids the technical problem in the prior art where environmental vibration and shaking of the detection equipment affect the measurement accuracy, resulting in low defect detection efficiency and limitations of the glass bead defect detection technology.

[0167] It should be noted that although the steps in the above embodiments are described in a specific order, those skilled in the art will understand that in order to achieve the effects of the present invention, different steps do not necessarily have to be executed in such an order. They can be executed simultaneously (in parallel) or in other orders, and these variations are all within the scope of protection of the present invention.

[0168] In one implementation, the specific function can be described in steps S101-S104.

[0169] See appendix Figure 3 , Figure 3 This is a schematic diagram of a glass bead defect detection device based on an optimized image analysis algorithm, according to an embodiment of the present invention. Figure 3 As shown, the glass bead defect detection device 1 based on optimized image analysis algorithm includes at least a feeding module 11, a good product unloading bin, a defective product unloading bin 12, a glass bead pushing module, and a glass bead detection module. The glass bead detection module is located around the glass bead pushing module. The feeding module 11 is used to perform the feeding operation in the glass bead detection process. The good product unloading bin is where the good product material is located, and the defective product unloading bin 12 is where the defective product material is located. The glass bead detection module includes at least a camera module and a control device. The number of camera modules is at least 8.

[0170] A glass bead defect detection device 1 based on an optimized image analysis algorithm is used for performing... Figure 1 The glass bead defect detection method based on the optimized image analysis algorithm shown in the embodiments are similar in technical principle, technical problem solved and technical effect produced. Those skilled in the art can clearly understand that, for the sake of convenience and brevity, the specific working process and related descriptions of the glass bead defect detection device 1 based on the optimized image analysis algorithm can be referred to the content described in the embodiments of the glass bead defect detection method based on the optimized image analysis algorithm, and will not be repeated here.

[0171] Those skilled in the art will understand that all or part of the processes in the method of the above-described embodiment of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or any intermediate form. The computer-readable storage medium can include any entity or device capable of carrying the computer program code, a medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content included in the computer-readable storage medium can be appropriately added to or subtracted according to the requirements of legislation and patent practice in any jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.

[0172] Furthermore, the control device of the glass bead defect detection device 1 based on the optimized image analysis algorithm of the present invention includes a processor and a storage device. The storage device can be configured to store a program for executing the glass bead defect detection method based on the optimized image analysis algorithm described in the above method embodiments. The processor can be configured to execute the program in the storage device, which includes, but is not limited to, the program for executing the glass bead defect detection method based on the optimized image analysis algorithm described in the above method embodiments. For ease of explanation, only the parts related to the embodiments of the present invention are shown. For specific technical details not disclosed, please refer to the method section of the embodiments of the present invention. The control device can be a control device device device comprising various electronic devices.

[0173] Furthermore, the glass bead defect detection device 1 based on the optimized image analysis algorithm of the present invention also includes a computer-readable storage medium. In one embodiment of the present invention, the computer-readable storage medium can be configured to store a program for executing the glass bead defect detection method based on the optimized image analysis algorithm described in the above-described method embodiments. This program can be loaded and run by a processor to implement the glass bead defect detection method based on the optimized image analysis algorithm described above. For ease of explanation, only the parts related to the embodiments of the present invention are shown; for specific technical details not disclosed, please refer to the method section of the embodiments of the present invention. The computer-readable storage medium can be a storage device comprising various electronic devices. Optionally, in the embodiments of the present invention, the computer-readable storage medium is a non-transitory computer-readable storage medium.

[0174] Furthermore, it should be understood that since the various modules are only provided to illustrate the functional units of the device of the present invention, the physical devices corresponding to these modules may be the processor itself, or a part of the processor's software, hardware, or a combination of software and hardware. Therefore, the number of modules shown in the figures is merely illustrative.

[0175] Those skilled in the art will understand that the various modules in the device can be adaptively split or combined. Such splitting or combining of specific modules will not cause the technical solution to deviate from the principles of the present invention; therefore, the technical solutions after splitting or combining will fall within the protection scope of the present invention.

[0176] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A method for detecting defects in glass beads based on an optimized image analysis algorithm, characterized in that, The method includes the following steps: Acquire multi-angle image data of the glass bead; Based on the multi-angle image data of the glass bead, a first defect detection is performed on the glass bead: If a defect is detected, the glass bead is marked as a defective glass bead. Otherwise, the multi-angle image data is substituted into the defect detection module, so that the glass bead is selectively inspected in two dimensions or three dimensions based on the multi-angle image data, and the detection result of the glass bead is obtained. The detection result includes whether the glass bead has a defect, and when the glass bead has a defect, the type of defect of the glass bead. Based on the test results of the glass beads, the glass beads are marked as defective glass beads.

2. The glass bead defect detection method based on optimized image analysis algorithm according to claim 1, characterized in that, The step of substituting the multi-angle image data into the defect detection module, so that the glass bead is selectively subjected to two-dimensional or three-dimensional detection based on the multi-angle image data, to obtain a detection result indicating whether the glass bead has defects includes: Based on the multi-angle image data, detect whether there are interference items in the multi-angle image data of the glass bead: If it is determined that the glass bead exists, the process of "acquiring multi-angle image data of the glass bead" is repeated, and the number of times the multi-angle image data of the glass bead is acquired, as well as subsequent steps, is recorded. If it is determined that the glass bead does not exist, or if the number of times the multi-angle image of the glass bead is acquired is 2, then the glass bead is fused based on all the multi-angle image data of the glass bead to obtain the fused image of the glass bead. Based on the fused image of the glass beads, two-dimensional or three-dimensional defect detection is selectively performed on the glass beads.

3. The glass bead defect detection method based on optimized image analysis algorithm according to claim 2, characterized in that, The fused image based on glass beads, selectively further performing two-dimensional or three-dimensional defect detection on the glass beads, includes: Based on the fused image of the glass beads, a second defect detection is performed on the glass beads: If a defect is detected, the defect type of the glass bead is determined, and a detection result indicating that the glass bead has a defect is generated, thereby achieving two-dimensional detection of the glass bead. If no defects are detected, a three-dimensional surface model of the glass beads is generated based on the fused image of the glass beads. Based on the three-dimensional surface model of the glass bead, determine whether the glass bead has defects: If the defect is determined to exist, the defect type of the glass bead is determined, and the detection result of the glass bead is generated to achieve three-dimensional detection of the glass bead. Otherwise, a test result indicating that the glass bead has no defects is generated.

4. The glass bead defect detection method based on optimized image analysis algorithm according to claim 3, characterized in that, The step of detecting whether there are interference items in the multi-angle image data of the glass bead based on the multi-angle image data includes: The degree of image interference can be determined using the following formula: ； in, Representative image The degree of interference, Represents the image spatial domain. Represents the image at location The Laplace operator, The mean of the Laplace operator, The weight coefficients representing feature point matching. as well as These represent the first and second images in the reference and test images, respectively. One feature point, The weighting coefficients representing the frequency domain influence of vibration. Represents frequency variables. Represents the Fourier transform of the image. Represents the parameters of the vibration transfer function; when When the multi-angle image data of the glass bead is detected, it is determined that there are interference items. This represents the preset threshold for the degree of interference, with a value of [value missing]. ; Otherwise, it is determined that there are no interference items in the multi-angle image data of the glass bead being detected.

5. The glass bead defect detection method based on optimized image analysis algorithm according to claim 4, characterized in that, The process of fusing all multi-angle image data of the glass bead to obtain a fused image of the glass bead includes: The fusion image of the glass beads is obtained using the following formula: ； in, Represents the merged image. The first representing glass beads The image of the frame, The weighting coefficients representing the smoothing term. The value ranges from [0.01, 1.0]. The weight coefficients representing the Huber loss function. The value ranges from [0.1, 1.0]. This represents the Huber loss function value.

6. The glass bead defect detection method based on optimized image analysis algorithm according to claim 5, characterized in that, The determination of whether the glass bead has defects based on the three-dimensional surface model of the glass bead includes: The following formula is used to monitor whether the glass beads have defects: ； in, This represents the energy functional for defect detection. Represents the defect probability field. A three-dimensional surface model representing a glass bead. Represents the gradient term The weighting coefficients, Representing Laplace's term The weighting coefficients, The weighting coefficients represent the boundary terms. Represents boundary conditions. Represents the three-dimensional volume region inside the bubble. Represents the surface boundary of the bubble; Defect detection is performed on the glass beads based on the defect detection energy functional.

7. The glass bead defect detection method based on optimized image analysis algorithm according to claim 6, characterized in that, The process of determining the defect type of the glass bead and generating the detection result of the glass bead to achieve three-dimensional detection of the glass bead includes: The defect type of the glass bead is determined by the following formula: ； in, Represents the defect type of the glass bead. The feature vector representing the defect to be tested. Representative characteristics template feature vector, Representing the Each feature similarity parameter as well as These represent the classifier parameters.

8. The glass bead defect detection method based on optimized image analysis algorithm according to claim 7, characterized in that, The feature vectors of the defects to be tested include at least bubble vectors, color vectors, protrusion and pit vectors, scar and white mark vectors, and wear vectors. The bubble vector in the feature vector of the defect to be tested is obtained by the following formula: ； in, Represents the bubble vector. Represents the curvature of the surface inside the bubble. Represents the bubble density field. Represents the divergence of the normal vector. Represents the boundary term weight coefficients; The heterochromatic vector in the feature vector of the defect to be tested is obtained by the following formula: ； in, Represents a heterochromatic vector. Represents the square of the color gradient magnitude. Represents the energy of the color spectrum. The Laplacian operator represents the color. as well as These represent the weighting coefficients.

9. The glass bead defect detection method based on optimized image analysis algorithm according to claim 8, characterized in that, The convex and concave vectors in the feature vector of the defect to be tested are obtained by the following formula: ； in, Represents the vectors of bulges and depressions. The curvature representing the protruding defect. The curvature representing a protruding defect is a positive value. The curvature representing a protruding defect is negative. The average value representing the background curvature; The scar, white mark, and wear vector in the feature vector of the defect to be tested are obtained by the following formula: ； in, This represents the squared magnitude of the color gradient, used to reflect color abrupt changes in scars / white marks. This represents the second-order color gradient, used to reflect the gradual color change in the wear area. Represents the curvature of the defective region. Represents the curvature gradient; Furthermore, through the matrix The combination of element values ​​determines the defect type of scars, white marks, and wear vectors: 1) Scars: High color gradient, high curvature, high curvature gradient; 2) White print: High color gradient, low curvature, low curvature gradient; 3) Wear: Low color gradient, low curvature, low curvature gradient.

10. The glass bead defect detection method based on optimized image analysis algorithm according to claim 9, characterized in that, The method further includes: During the detection process of the defect detection module, a comprehensive score is given for the detection process of any number of glass beads randomly selected, to ensure the accuracy of the defect detection module and to optimize the defect detection module in a timely manner. The comprehensive score for the glass bead testing process is obtained using the following formula: ； in, The index represents the evaluation index of fusion effect. This represents the index for evaluating the effectiveness of defect detection. , , , All represent weighting coefficients, satisfying , This represents the gradient field of the overall score, used to constrain the smoothness of the score.

Images