Injection molding part surface treatment effect quality inspection method and device based on machine vision

Abstract

The invention discloses an injection molding part surface treatment effect quality inspection method and device based on machine vision, and relates to the technical field of injection molding part surface treatment effect visual quality inspection. Detection light sources with the same angle and intensity are adopted for visual picture shooting, a visual analysis module carries out double-display-effect visual classification on an image to be subjected to quality inspection based on a machine learning algorithm, and the double-display-effect visual classification is that each display effect is independently classified; performing independent classification result combination based on each display effect to realize classification of visual effects; the uniformity of dissimilar parts is considered in display effect evaluation, and evaluation of the uniformity of the dissimilar parts is added, so that the accuracy of visual quality inspection can be improved; intelligent quality inspection of the surface visual effect performance of the injection molding part with the double-display effect is achieved, subjectivity of manual quality inspection is avoided, and full inspection matched with the takt of a production line can be achieved.

Description

Technical Field

[0001] The invention relates to the technical field of visual quality inspection of surface treatment effects of injection molded parts, and in particular to a method and device for quality inspection of surface treatment effects of injection molded parts based on machine vision. Background Art

[0002] With the diversification and personalization of automotive interior and exterior trims, automotive exterior trims have been given more changes. The PVD coating process is added to the exterior trim to increase the sense of color layering. For existing processes, the material can generally only present a single color state after painting. There are also some special injection molding surface treatment processes that can achieve double-surface visual effects, such as electroplating and high-gloss black paint effects on the surface of the injection molding, or electroplating and light-transmitting effects on the surface of the injection molding.

[0003] In conventional injection molding production lines, visual methods are generally used to detect surface dimensions and defects. The quality inspection of the visual effects presented on the surface is often based on empirical judgments made through manual sampling. For example, whether the visual effects meet the requirements is judged based on experience at a distance of about 0.5 meters.

[0004] The Chinese invention patent application with publication number CN108830832A discloses a plastic tube surface defect detection algorithm based on machine vision, which includes the following steps: a line scan camera visual imaging system acquires images; the image is enhanced by a piecewise linear transformation grayscale transformation method; a median filter is used to denoise the region; the image is segmented using a LOG operator to obtain printed edges; the image region ROI is extracted from the surface of the plastic tube; the image is segmented using a threshold method to obtain the connected region of the segmented image; and defects are determined based on the area and edge features of the connected region.

[0005] The above visual inspection method is mainly aimed at detecting impurities on the surface of plastic cylinders, and cannot achieve quality inspection of visual effects;

[0006] Therefore, there is an urgent need to develop visual quality inspection methods for the complex visual effects of large quantities of automotive interior and exterior trims. Summary of the invention

[0007] In order to solve the above-mentioned technical problem of visual quality inspection of injection molded parts surface, the present invention provides a method and device for quality inspection of injection molded parts surface treatment effect based on machine vision. The following technical solution is adopted:

[0008] The quality inspection method of the surface treatment effect of injection molded parts based on machine vision includes the following steps:

[0009] Step 1: The transfer robot places the injection molded parts to be inspected on the visual inspection tooling;

[0010] Step 2: The detection light source is turned on, and the visual camera takes visual images of the injection molded parts to be inspected from multiple angles. The visual images are numbered based on the numbers of the injection molded parts to be inspected, and the visual images with numbering data are transmitted to the visual analysis module;

[0011] Step 3: The visual analysis module pre-processes the visual image to obtain the image to be inspected;

[0012] Step 4: The visual analysis module performs dual display effect visual classification on the image to be inspected based on a machine learning algorithm, and grades the visual effects based on the dual display effect visual classification results, and stores the visual classification results, the grading results, and the image to be inspected based on the numbering data;

[0013] Step 5: The transfer robot packs the injection molded parts to be inspected into the qualified product storage rack, the processed product storage rack, and the unqualified product storage rack according to the visual grading results of the dual display effects.

[0014] By adopting the above technical solution, a visual inspection tooling capable of supporting the injection molded parts to be inspected is used as the standard position for visual acquisition, and a detection light source with the same angle and intensity is used to shoot the visual picture. The visual analysis module performs dual display effect visual classification of the inspection image based on the machine learning algorithm. The visual analysis module refers to a device with visual analysis computing capabilities, such as a computer deployed with a machine learning algorithm model. The collected visual picture is pre-processed to obtain the inspection image, and then the inspection image is input into the machine learning algorithm model. The machine learning algorithm model is trained with a large number of standard display effect images corresponding to the dual display effects. For example, for a monochrome injection molded part, a visual image with dual display effects of electroplating and high-gloss black paint is obtained. During quality inspection, a large number of visual images containing dual display effects of standard electroplating and high-gloss black paint are used to train the model. Finally, the machine learning algorithm will perform dual display effect visual classification on the images to be inspected. The dual display effect visual classification is to classify each display effect separately, and then the visual effects are graded based on the combination of the results of the separate classification of each display effect. Finally, the transfer robot divides the injection molded parts to be inspected into qualified product storage racks, pending product storage racks and unqualified product storage racks according to the dual display effect visual grading results, realizing intelligent quality inspection of the surface visual effects of dual display effect injection molded parts, improving quality inspection efficiency and accuracy, avoiding the subjectivity of manual quality inspection, and achieving full inspection that matches the production line rhythm.

[0015] Optionally, in step 3, the method in which the visual analysis module preprocesses the visual image includes the following sub-steps:

[0016] Step 31, performing filtering, denoising, and contrast enhancement operations;

[0017] Step 32, using an edge detection algorithm, firstly separates the background edge of the injection molded part, a plurality of first display effect parts, and a plurality of second display effect parts.

[0018] By adopting the above technical solution, after filtering, denoising and contrast enhancement operations, the influence of environmental factors on visual effect recognition can be reduced;

[0019] Use an edge detection algorithm to separate the background edge of the injection molded part, and then further separate multiple first display effect parts and multiple second display effect parts. For example, in the visual quality inspection of monochrome injection molded parts showing dual display effects of electroplating and high-gloss black paint, multiple electroplating display effect parts and multiple bright black paint display effect parts can be separated, and the similarity of a single display effect can be compared with the corresponding standard display effect.

[0020] Optionally, a color segmentation technology is used to distinguish specific display colors of the multiple first display effect parts and the multiple second display effect parts.

[0021] By adopting the above technical solution, usually there should be a color difference between the two display effects of the dual display effect, and the first display effect part and the second display effect part can be distinguished by using the color segmentation technology.

[0022] Optionally, in step 4, the visual analysis module classifies the first display effect based on the machine learning algorithm into Aa, Ab, Ac and Ad, and the similarity between the first display effect and the standard first display effect is X, wherein Aa represents X ≥ 98%, Ab represents 95% ≤ X < 98%, Ac represents 90% ≤ X < 95%, and Ad represents X < 90%;

[0023] The classification results of the second display effect by the machine learning algorithm include Ba, Bb, Bc and Bd. Assume that the similarity between the second display effect and the standard second display effect is Y, where Ba represents Y ≥ 98%, Bb represents 95% ≤ Y < 98%, Bc represents 90% ≤ Y < 95%, and Bd represents Y < 90%;

[0024] If the first display effect and the second display effect combination of the dual display effect visual classification are classified as AaBa, AaBb or AbBa, the visual effect grading result is qualified;

[0025] If the first display effect and the second display effect combination of the dual display effect visual classification are classified as AbBb, AaBc or AcBa, the visual effect classification result is pending;

[0026] If the dual display effect visual classification is other combinations, the visual effect grading result will be unqualified.

[0027] By adopting the above technical solution, visual effect classification is implemented by single display effect classification, wherein the classification is implemented based on the similarity analysis result. For a display effect, if its similarity with the standard display effect is greater than 98%, it means that it is very close to the standard display effect. If the similarity is between 95% and 98%, it means that although there is a slight difference, the visual presentation can also meet the requirements. Similarly, if the similarity is less than 90%, it is considered that the visual effect is greatly different from the standard visual effect, and it is unqualified for monochrome.

[0028] In the final classification, the combined results of the two display effects were analyzed. For the classification combination of AaBa, the dual display effect was the best, and AaBb or AbBa was visually acceptable, so it was judged as qualified;

[0029] For products classified as AbBb, AaBc or AcBa, if they are manually observed and have minor visual defects, they need to be manually re-inspected. If some minor defects can be repaired, such as erasing paint spots and stains, and can pass manual quality inspection, they can be manually repaired and inspected and judged as qualified products;

[0030] For other combination classifications, they can be directly judged as unqualified products.

[0031] Thus, the automatic visual grading of dual display effects is quantitatively achieved.

[0032] Optionally, in step 4, the similarity between the display effect and the standard display effect is calculated using the following method:

[0033] Extract features from the image to be inspected and the standard surface image to form feature vectors AA and BB;

[0034] Calculate the dot product of feature vectors AA and BB, calculate the Euclidean norms of feature vectors AA and BB respectively, divide the dot product by the product of the two norms to get the cosine similarity value;

[0035] The value range of cosine similarity is between [0%, 100%], where 100% means that the two vectors are exactly the same, and 0 means that the two vectors are orthogonal, that is, they are not similar;

[0036] The cosine similarity value is the similarity value between the display effect and the standard display effect.

[0037] Optionally, the cosine similarity between the display effect and the standard display effect is calculated using the following formula:

[0038] ;

[0039] in Represents the display effect feature vector, represents the standard display effect feature vector, Represents the cosine similarity value between the display effect vector and the standard display effect vector, Representation vector and vector The dot product of Represents vectors and vector The Euclidean norm of .

[0040] By adopting the above technical solution, the value range of cosine similarity is between [-1, 1], where 1 means that the two vectors are exactly the same, 0 means that the two vectors are orthogonal (i.e., not similar), and -1 means that the two vectors are completely opposite. In practical applications, only non-negative similarities are usually considered, so the value of cosine similarity is usually in the range of [0, 1]. In this way, the similarity between the photographed injection molded part surface and the standard surface can be quantitatively evaluated, thereby performing quality inspection.

[0041] Optionally, after calculating the cosine similarity value, the uniformity U of the dissimilar part is calculated, and a first uniformity threshold Ua and a second uniformity threshold Ub are set. If U≥Ua, the similarity value between the display effect and the standard display effect is output based on the cosine similarity value. If Ub≤U<Ua, the cosine similarity value is subtracted by 3% and output as the similarity value between the display effect and the standard display effect; if U<Ub, the similarity value between the display effect and the standard display effect is directly output as less than 90%.

[0042] By adopting the above technical solution, the uniformity of the dissimilar parts needs to be considered in the display effect evaluation. In layman's terms, if the dissimilar parts are uniform, then they will not have too much impact on the visual effect. If the dissimilar parts are in a clustered state, even if the dissimilar parts only account for 2%, it may cause visual defects visible to the naked eye. Therefore, a two-stage uniformity threshold is set. For example, the uniformity is in the range of [0, 1], 1 is the most uniform, and the first uniformity threshold Ua=0.9 and the second uniformity threshold Ub=0.85 are set. If the uniformity of the dissimilar parts U≥0.9, it is considered that the uniformity of the dissimilar parts meets the requirements and does not affect the visual display effect. If 0.85≤U<0.9, it is considered that there is an impact. At this time, the cosine similarity value is subtracted by 3% and output as the similarity value between the display effect and the standard display effect, and the display effect is degraded and evaluated. If U<0.85, it is considered that there are clustered defects. At this time, it is necessary to directly judge that the visual effect is unqualified. Adding the evaluation of the uniformity of the dissimilar parts can improve the accuracy of the visual quality inspection.

[0043] Optionally, the uniformity of the dissimilar parts can be calculated using the following method: :

[0044] Calculate the difference vector between two vectors in the direction of dissimilarity ;

[0045] ;

[0046] Calculate the standard deviation of the difference vector :

[0047] ;

[0048] in is the difference vector Quantity, is the dimension of the vector, is the mean of the difference vectors;

[0049] Calculate the uniformity of dissimilar parts :

[0050] ;

[0051] in is a constant.

[0052] By adopting the above technical solution, the uniformity of the dissimilar parts can be quantitatively calculated. .

[0053] An injection molded part surface treatment effect quality inspection device based on machine vision comprises a visual quality inspection tool, a visual camera, a visual analysis module and a storage rack group, wherein the visual quality inspection tool is provided with a plurality of adjustable support blocks and a detection light source, and when a transfer robot places the mounting surface of the injection molded part to be inspected on the upper surface of the plurality of adjustable support blocks, the surface of the injection molded part to be inspected faces upward, and after the transfer robot moves away, the detection light source is turned on, and the visual camera captures a surface visual image of the injection molded part to be inspected, and communicates with the visual analysis module to exchange the surface visual image, and the visual analysis module runs a quality inspection program designed using the injection molded part surface treatment effect quality inspection method based on machine vision as described in claim 8, outputs quality inspection results and stores the quality inspection results according to the numbers of the injection molded parts to be inspected.

[0054] Optionally, the storage rack group includes a storage rack for qualified products, a storage rack for products to be processed, and a storage rack for unqualified products. The transfer robot communicates with the visual analysis module to exchange quality inspection results, and based on the quality inspection results, transfers the injection molded parts that have completed quality inspection to one of the storage racks for qualified products, the storage racks for products to be processed, and the storage racks for unqualified products.

[0055] In summary, the present invention includes at least one of the following beneficial technical effects:

[0056] The present invention can provide a method and device for quality inspection of surface treatment effects of injection molded parts based on machine vision, adopt a visual quality inspection tooling that can support the injection molded parts to be inspected as a standard position for visual acquisition, adopt a detection light source with the same angle and intensity to shoot visual images, and a visual analysis module performs dual display effect visual classification on the inspection image based on a machine learning algorithm. The dual display effect visual classification is to classify each display effect separately, and then to classify the visual effects based on the combination of the results of the separate classification of each display effect. Finally, the transfer robot packs the injection molded parts to be inspected into a qualified product storage rack, a storage rack for products to be processed, and a storage rack for unqualified products according to the dual display effect visual grading results.

[0057] Considering the uniformity of dissimilar parts in the display effect evaluation, adding the evaluation of the uniformity of dissimilar parts can improve the accuracy of visual quality inspection;

[0058] It realizes intelligent quality inspection of the surface visual effects of injection molded parts with dual display effects, improves quality inspection efficiency and accuracy, avoids the subjectivity of manual quality inspection, and can achieve full inspection that matches the rhythm of the production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a flow chart of the quality inspection method of the surface treatment effect of injection molded parts based on machine vision of the present invention;

[0060] Figure 2 It is a visual image to be inspected of the dual display effects of electroplating and high-gloss black paint of the present invention. DETAILED DESCRIPTION

[0061] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0062] The embodiment of the present invention discloses a method and a device for quality inspection of surface treatment effects of injection molded parts based on machine vision.

[0063] Reference Figure 1 and Figure 2 Embodiment 1, a method for inspecting the surface treatment effect of injection molded parts based on machine vision, comprising the following steps:

[0064] Step 1: The transfer robot places the injection molded parts to be inspected on the visual inspection tooling;

[0065] Step 2: The detection light source is turned on, and the visual camera takes visual images of the injection molded parts to be inspected from multiple angles. The visual images are numbered based on the numbers of the injection molded parts to be inspected, and the visual images with numbering data are transmitted to the visual analysis module;

[0066] Step 3: The visual analysis module pre-processes the visual image to obtain the image to be inspected;

[0067] Step 4: The visual analysis module performs dual display effect visual classification on the image to be inspected based on a machine learning algorithm, and grades the visual effects based on the dual display effect visual classification results, and stores the visual classification results, the grading results, and the image to be inspected based on the numbering data;

[0068] Step 5: The transfer robot packs the injection molded parts to be inspected into the qualified product storage rack, the processed product storage rack, and the unqualified product storage rack according to the visual grading results of the dual display effects.

[0069] A visual inspection tooling that can support the injection molded parts to be inspected is used as the standard position for visual acquisition, and a detection light source with the same angle and intensity is used to capture the visual image. The visual analysis module performs dual display effect visual classification on the inspection image based on the machine learning algorithm. The visual analysis module refers to a device with visual analysis computing capabilities, such as a computer that deploys a machine learning algorithm model. The collected visual images are pre-processed to obtain the inspection image, and then the inspection image is input into the machine learning algorithm model. The machine learning algorithm model is trained with a large number of standard display effect images corresponding to the dual display effect. For example, in the visual inspection of monochrome injection molded parts with dual display effects of electroplating and high-gloss black paint, see Figure 2 A large number of visual images containing dual display effects of standard electroplating and high-gloss black paint are used to train the model. Finally, the machine learning algorithm will perform dual display effect visual classification on the images to be inspected. The dual display effect visual classification is to classify each display effect separately, and then the visual effects are graded based on the combination of the results of the separate classification of each display effect. Finally, the transfer robot divides the injection molded parts to be inspected into qualified product storage racks, pending product storage racks and unqualified product storage racks according to the dual display effect visual grading results, realizing intelligent quality inspection of the surface visual effects of dual display effect injection molded parts, improving quality inspection efficiency and accuracy, avoiding the subjectivity of manual quality inspection, and achieving full inspection that matches the production line rhythm.

[0070] In step 3 of embodiment 2, the method for preprocessing the visual image by the visual analysis module includes the following sub-steps:

[0071] Step 31, performing filtering, denoising, and contrast enhancement operations;

[0072] Step 32, using an edge detection algorithm, firstly separates the background edge of the injection molded part, a plurality of first display effect parts, and a plurality of second display effect parts.

[0073] After filtering, denoising, and contrast enhancement operations, the impact of environmental factors on visual effect recognition can be reduced;

[0074] Use an edge detection algorithm to separate the background edge of the injection molded part, and then further separate multiple first display effect parts and multiple second display effect parts. For example, in the visual quality inspection of monochrome injection molded parts showing dual display effects of electroplating and high-gloss black paint, multiple electroplating display effect parts and multiple bright black paint display effect parts can be separated, and the similarity of a single display effect can be compared with the corresponding standard display effect.

[0075] In the third embodiment, a color segmentation technology is used to distinguish specific display colors of a plurality of first display effect parts and a plurality of second display effect parts.

[0076] Usually, there should be a color difference between the two display effects of the dual display effect, and the color segmentation technology can be used to distinguish the first display effect part and the second display effect part.

[0077] In Example 4, in step 4, the visual analysis module classifies the first display effect based on the machine learning algorithm into Aa, Ab, Ac and Ad. Assume that the similarity between the first display effect and the standard first display effect is X, where Aa represents X ≥ 98%, Ab represents 95% ≤ X < 98%, Ac represents 90% ≤ X < 95%, and Ad represents X < 90%;

[0078] The classification results of the second display effect by the machine learning algorithm include Ba, Bb, Bc and Bd. Assume that the similarity between the second display effect and the standard second display effect is Y, where Ba represents Y ≥ 98%, Bb represents 95% ≤ Y < 98%, Bc represents 90% ≤ Y < 95%, and Bd represents Y < 90%;

[0079] If the first display effect and the second display effect combination of the dual display effect visual classification are classified as AaBa, AaBb or AbBa, the visual effect grading result is qualified;

[0080] If the first display effect and the second display effect combination of the dual display effect visual classification are classified as AbBb, AaBc or AcBa, the visual effect classification result is pending;

[0081] If the dual display effect visual classification is other combinations, the visual effect grading result will be unqualified.

[0082] Visual effect classification is implemented by single display effect classification, where classification is implemented based on similarity analysis results. For a display effect, if its similarity with the standard display effect is greater than 98%, it means that it is very close to the standard display effect. If the similarity is between 95% and 98%, it means that although there is a slight difference, the visual presentation can still meet the requirements. Similarly, if the similarity is less than 90%, it is considered that the visual effect is greatly different from the standard visual effect and is unqualified for monochrome.

[0083] In the final classification, the combined results of the two display effects were analyzed. For the classification combination of AaBa, the dual display effect was the best, and AaBb or AbBa was visually acceptable, so it was judged as qualified;

[0084] For products classified as AbBb, AaBc or AcBa, if they are manually observed and have minor visual defects, they need to be manually re-inspected. If some minor defects can be repaired, such as erasing paint spots and stains, and can pass manual quality inspection, they can be manually repaired and inspected and judged as qualified products;

[0085] For other combination classifications, they can be directly judged as unqualified products.

[0086] Thus, the automatic visual grading of dual display effects is quantitatively achieved.

[0087] In Example 5, in step 4, the similarity between the display effect and the standard display effect is calculated using the following method:

[0088] Extract features from the image to be inspected and the standard surface image to form feature vectors AA and BB;

[0089] Calculate the dot product of feature vectors AA and BB, calculate the Euclidean norms of feature vectors AA and BB respectively, divide the dot product by the product of the two norms to get the cosine similarity value;

[0090] The value range of cosine similarity is between [0%, 100%], where 100% means that the two vectors are exactly the same, and 0 means that the two vectors are orthogonal, that is, they are not similar;

[0091] The cosine similarity value is the similarity value between the display effect and the standard display effect.

[0092] In Example 6, the cosine similarity between the display effect and the standard display effect is calculated using the following formula:

[0093] ;

[0094] in Represents the display effect feature vector, represents the standard display effect feature vector, Represents the cosine similarity value between the display effect vector and the standard display effect vector, Representation vector and vector The dot product of Represents vectors and vector The Euclidean norm of .

[0095] The value range of cosine similarity is between [-1, 1], where 1 means that the two vectors are exactly the same, 0 means that the two vectors are orthogonal (i.e., not similar), and -1 means that the two vectors are completely opposite. In practical applications, only non-negative similarities are usually considered, so the value of cosine similarity is usually in the range of [0, 1]. In this way, the similarity between the photographed injection molded part surface and the standard surface can be quantitatively evaluated for quality inspection.

[0096] Embodiment 7, after calculating the cosine similarity value, calculate the uniformity U of the dissimilar part, set the first uniformity threshold Ua and the second uniformity threshold Ub, if U≥Ua, then based on the cosine similarity value, output as the similarity value between the display effect and the standard display effect, if Ub≤U<Ua, then subtract 3% from the cosine similarity value and output as the similarity value between the display effect and the standard display effect; if U<Ub, then the similarity value between the display effect and the standard display effect is directly output as less than 90%.

[0097] In the evaluation of display effect, the uniformity of dissimilar parts needs to be considered. In layman's terms, if dissimilar parts are uniform, they will not have too much impact on the visual effect. If dissimilar parts are in a clustered state, even if the dissimilar parts only account for 2%, they may cause visual defects visible to the naked eye. Therefore, a two-stage uniformity threshold is set. For example, the uniformity is in the range of [0, 1], 1 is the most uniform, and the first uniformity threshold Ua=0.9 and the second uniformity threshold Ub=0.85 are set. If the uniformity of the dissimilar parts U≥0.9, it is considered that the uniformity of the dissimilar parts meets the requirements and does not affect the visual display effect. If 0.85≤U<0.9, it is considered that there is an impact. At this time, the cosine similarity value is subtracted by 3% and output as the similarity value between the display effect and the standard display effect. The display effect is degraded and evaluated. If U<0.85, it is considered that there are clustered defects. At this time, it is necessary to directly judge that the visual effect is unqualified. Adding the evaluation of the uniformity of dissimilar parts can improve the accuracy of visual quality inspection.

[0098] Example 8, the uniformity of the dissimilar parts is calculated using the following method :

[0099] Calculate the difference vector between two vectors in the direction of dissimilarity ;

[0100] ;

[0101] Calculate the standard deviation of the difference vector :

[0102] ;

[0103] in is the difference vector Quantity, is the dimension of the vector, is the mean of the difference vectors;

[0104] Calculate the uniformity of dissimilar parts :

[0105] ;

[0106] in is a constant.

[0107] The uniformity of dissimilar parts can be quantitatively calculated .

[0108] Embodiment 9, a device for inspecting the surface treatment effect of injection molded parts based on machine vision, comprises a visual inspection tool, a visual camera, a visual analysis module and a storage rack group, wherein the visual inspection tool is provided with a plurality of adjustable support blocks and a detection light source, and when the transfer robot places the mounting surface of the injection molded part to be inspected on the upper surface of the plurality of adjustable support blocks, the surface of the injection molded part to be inspected faces upward, and after the transfer robot moves away, the detection light source is turned on, and the visual camera takes a surface visual image of the injection molded part to be inspected, and communicates with the visual analysis module to exchange the surface visual image, and the visual analysis module runs a quality inspection program designed using the method for inspecting the surface treatment effect of injection molded parts based on machine vision according to claim 8, outputs the quality inspection results and stores the quality inspection results according to the numbers of the injection molded parts to be inspected.

[0109] In embodiment 10, the storage rack group includes a storage rack for qualified products, a storage rack for products to be processed, and a storage rack for unqualified products. The transfer robot communicates with the visual analysis module to exchange quality inspection results, and based on the quality inspection results, transfers the injection molded parts that have completed quality inspection to one of the storage racks for qualified products, the storage racks for products to be processed, and the storage racks for unqualified products.

[0110] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for quality inspection of surface treatment effects of injection molded parts based on machine vision, characterized in that: The following steps are involved: Step 1: The transfer robot places the injection molded parts to be inspected on the visual inspection tooling; Step 2: The detection light source is turned on, and the visual camera takes visual images of the injection molded parts to be inspected from multiple angles. The visual images are numbered based on the numbers of the injection molded parts to be inspected, and the visual images with numbering data are transmitted to the visual analysis module; Step 3: The visual analysis module pre-processes the visual image to obtain the image to be inspected; Step 4: The visual analysis module performs dual display effect visual classification on the image to be inspected based on a machine learning algorithm, and grades the visual effects based on the dual display effect visual classification results, and stores the visual classification results, the grading results, and the image to be inspected based on the numbering data; Step 5: The transfer robot packs the injection molded parts to be inspected into the qualified product storage rack, the processed product storage rack, and the unqualified product storage rack according to the visual grading results of the dual display effects.

2. The method for quality inspection of surface treatment effect of injection molded parts based on machine vision according to claim 1, characterized in that: In step 3, the method of preprocessing the visual image by the visual analysis module includes the following sub-steps: Step 31, performing filtering, denoising, and contrast enhancement operations; Step 32, using an edge detection algorithm, firstly separates the background edge of the injection molded part, a plurality of first display effect parts, and a plurality of second display effect parts.

3. The method for quality inspection of surface treatment effect of injection molded parts based on machine vision according to claim 2, characterized in that: The color segmentation technology is used to distinguish the specific display colors of the multiple first display effect parts and the multiple second display effect parts.

4. The method for quality inspection of surface treatment effect of injection molded parts based on machine vision according to claim 3, characterized in that: In step 4, the visual analysis module classifies the first display effect into Aa, Ab, Ac and Ad based on the machine learning algorithm. Let the similarity between the first display effect and the standard first display effect be X, where Aa represents X ≥ 98%, Ab represents 95% ≤ X < 98%, Ac represents 90% ≤ X < 95%, and Ad represents X < 90%; The classification results of the second display effect by the machine learning algorithm include Ba, Bb, Bc and Bd. Assume that the similarity between the second display effect and the standard second display effect is Y, where Ba represents Y ≥ 98%, Bb represents 95% ≤ Y < 98%, Bc represents 90% ≤ Y < 95%, and Bd represents Y < 90%; If the first display effect and the second display effect combination of the dual display effect visual classification are classified as AaBa, AaBb or AbBa, the visual effect grading result is qualified; If the first display effect and the second display effect combination of the dual display effect visual classification are classified as AbBb, AaBc or AcBa, the visual effect classification result is pending; If the dual display effect visual classification is other combinations, the visual effect grading result will be unqualified.

5. The method for quality inspection of surface treatment effect of injection molded parts based on machine vision according to claim 4, characterized in that: In step 4, the similarity between the display effect and the standard display effect is calculated using the following method: Extract features from the image to be inspected and the standard surface image to form feature vectors AA and BB; Calculate the dot product of feature vectors AA and BB, calculate the Euclidean norms of feature vectors AA and BB respectively, divide the dot product by the product of the two norms to get the cosine similarity value; The value range of cosine similarity is between [0%, 100%], where 100% means that the two vectors are exactly the same, and 0 means that the two vectors are orthogonal, that is, they are not similar; The cosine similarity value is the similarity value between the display effect and the standard display effect.

6. The method for quality inspection of surface treatment effect of injection molded parts based on machine vision according to claim 5, characterized in that: The cosine similarity between the display effect and the standard display effect is calculated using the following formula: ； in Represents the display effect feature vector, represents the standard display effect feature vector, Represents the cosine similarity value between the display effect vector and the standard display effect vector, Representation vector and vector The dot product of Represents vectors and vector The Euclidean norm of .

7. The method for quality inspection of surface treatment effect of injection molded parts based on machine vision according to claim 6, characterized in that: After calculating the cosine similarity value, the uniformity U of the dissimilar part is calculated, and the first uniformity threshold Ua and the second uniformity threshold Ub are set. If U≥Ua, the similarity value between the display effect and the standard display effect is output based on the cosine similarity value. If Ub≤U<Ua, the cosine similarity value is subtracted by 3% and output as the similarity value between the display effect and the standard display effect; if U<Ub, the similarity value between the display effect and the standard display effect is directly output as less than 90%.

8. The method for quality inspection of surface treatment effect of injection molded parts based on machine vision according to claim 7, characterized in that: The uniformity of the dissimilar parts is calculated using the following method : Calculate the difference vector between two vectors in the direction of dissimilarity ; ； Calculate the standard deviation of the difference vector : ； in is the difference vector Quantity, is the dimension of the vector, is the mean of the difference vector; Calculate the uniformity of dissimilar parts : ； in is a constant.

9. The device for inspecting the surface treatment effect of injection molded parts based on machine vision is characterized by: It includes a visual quality inspection tool, a visual camera, a visual analysis module and a storage rack group. The visual quality inspection tool is provided with a plurality of adjustable support blocks and a detection light source. When the transfer robot places the mounting surface of the injection molded part to be inspected on the upper surface of the plurality of adjustable support blocks, the surface of the injection molded part to be inspected faces upward. After the transfer robot moves away, the detection light source is turned on, and the visual camera takes a surface visual image of the injection molded part to be inspected, and communicates with the visual analysis module to exchange the surface visual image. The visual analysis module runs a quality inspection program designed using the injection molded part surface treatment effect quality inspection method based on machine vision as described in claim 8, outputs the quality inspection results and stores the quality inspection results according to the number of the injection molded part to be inspected.

10. The device for inspecting the surface treatment effect of injection molded parts based on machine vision according to claim 9, characterized in that: The storage rack group includes a storage rack for qualified products, a storage rack for products to be processed, and a storage rack for unqualified products. The transfer robot communicates with the visual analysis module to exchange quality inspection results, and based on the quality inspection results, transfers the injection molded parts that have completed quality inspection to one of the storage racks for qualified products, the storage racks for products to be processed, and the storage racks for unqualified products.

Images