A method for detecting thermal oxidation aging of automobile plastic parts

Abstract

The present invention relates to the technical field of material detection, and specifically to a method for detecting thermal oxygen aging of automobile plastic parts, the method comprising: using an industrial camera to photograph the surface of automobile plastic parts to obtain image data of the surface of automobile plastic parts; performing data preprocessing on the obtained image data of the surface of automobile plastic parts, including cropping, color correction and image enhancement; after the data preprocessing is completed, performing feature extraction on the color features and shape features of the image data of the surface of automobile plastic parts; after the feature extraction is completed, constructing and training a thermal oxygen aging detection model for automobile plastic parts based on an improved convolutional neural network; after the model training is completed, performing thermal oxygen aging detection on automobile plastic parts and outputting the detection results. The present invention captures the features of thermal oxygen aging of automobile plastic parts by combining image detection with a convolutional neural network method, accurately detects the thermal oxygen aging of automobile plastic parts, and improves the efficiency and automation level of detection.

Description

Technical Field

[0001] The invention relates to the technical field of magnetic material detection, and in particular to a method and system for detecting thermal oxygen aging of automobile plastic parts. Background Art

[0002] With the rapid development of the automotive industry, automotive plastic parts are increasingly used, playing an important role in reducing vehicle weight, reducing costs, and improving design flexibility. However, the use environment of automobiles makes plastic parts face severe challenges of thermal oxidative aging.

[0003] During the operation of the car, thermal factors such as solar radiation, ambient temperature changes, and heat generated by the engine, together with oxygen in the air, act on plastic parts. In the early stages of thermal oxygen aging, the color of plastic parts usually undergoes subtle changes that are not easily noticed, such as white plastic may turn slightly yellow, and black plastic may lose gloss. Over time, its mechanical properties gradually decrease, and its tensile strength decreases, making it easier to damage plastic parts when subjected to normal tensile forces; in addition, it also causes the impact strength of plastic parts to weaken, and they may break in the event of a minor collision. At the same time, surface roughness will also increase due to aging, affecting aesthetics and matching between components. In severe cases, plastic parts will have obvious cracks, deformation, or even structural collapse, threatening the safety and performance of the car.

[0004] The traditional detection method is manual visual inspection, which relies on the experience of the inspectors, is highly subjective and has low accuracy. It is difficult to detect early mild aging, and the detection efficiency of complex plastic parts is low. Although physical performance testing can accurately assess the degree of aging, it is destructive and requires sampling and testing with professional equipment. The process is cumbersome and time-consuming, and cannot meet the needs of large-scale automobile production and testing. Therefore, designing an efficient, non-destructive and accurate thermal oxygen aging detection method for automotive plastic parts has become an urgent problem to be solved in the industry. Summary of the invention

[0005] In view of the above-mentioned technical deficiencies, the present invention provides a method and system for detecting thermal oxidation aging of automobile plastic parts.

[0006] The present invention is achieved through the following technical solutions:

[0007] A method for detecting thermal oxidation aging of automobile plastic parts is provided, the method comprising the following steps:

[0008] Step S10: photographing the surface of the automobile plastic part with an industrial camera having high resolution and high color reproduction to obtain surface image data of the automobile plastic part, and ensuring uniform and stable lighting during the acquisition process to avoid the influence of shadows and reflections on the image quality;

[0009] Step S20: performing data preprocessing on the obtained automobile plastic part surface image data, including cropping, color correction and image enhancement;

[0010] Step S30: after the data preprocessing is completed, feature extraction is performed on the color features and shape features of the surface image data of the automobile plastic part;

[0011] Step S40: after feature extraction is completed, a thermal oxidation aging detection model for automobile plastic parts based on an improved convolutional neural network is constructed and trained;

[0012] Step S50: After the model training is completed, thermal oxygen aging test of automobile plastic parts is performed and the test results are output;

[0013] The industrial camera used in step S10 has a 2k resolution, and a ring light source is installed above the automobile plastic part to keep the lighting uniform and stable during the shooting process;

[0014] The data preprocessing step in step S20 includes cropping, color correction and image enhancement. Cropping is used to remove the background, the shooting table and the surrounding objects in the surface image of the automobile plastic part to reduce the amount of data processing; color correction is used to restore the true value of the image data, and the correction formula can be calibrated according to the standard color card; image enhancement is used to remove noise in the surface image of the automobile plastic part and enhance the recognition of aging features in the image;

[0015] The color extraction in step S30 is HSV color feature extraction, including hue, saturation and brightness; the shape feature extraction is to obtain the edge contour of the plastic part by cropping the surface image of the automobile plastic part and calculate the perimeter and area of ​​the edge contour;

[0016] Among them, the improved convolutional neural network in step S40 adopts a multi-level convolutional network architecture, 64 3×3 convolution kernels are added to the first convolution layer to extract local detail features of the surface image of automobile plastic parts, and a batch normalization layer is introduced after each convolution layer to standardize the data; in addition, residual connection is used in the network structure to construct a residual block, each residual block contains 2 convolution layers, the convolution kernel sizes are 3×3 and 1×1, and the number of convolution kernels is 128 and 64, respectively. A batch normalization layer is added between the two convolution layers, and ReLU is used as the activation function; the fully connected layer is set to 3 layers, and the number of neurons in each layer is 256, 128 and 64, respectively. The neuron connections are discarded according to the set ratio between the fully connected layers through the Dropout method to prevent overfitting. The output of the last layer is the classification result of thermal oxidation aging of automobile plastic parts. The Softmax function is used for multi-classification output, and the output is converted into the probability distribution of each category.

[0017] Preferably, the step S20 performs data preprocessing on the obtained automobile plastic part surface image data, including:

[0018] Step S201: According to the position and outline of the automobile plastic part in the image, the boundary coordinates are determined, the shape of the edge outline of the automobile plastic part is obtained by the Canny edge detection algorithm, and the cropping area is set according to the obtained boundary coordinates and edge outline, and the surface image of the automobile plastic part is cropped by using the cropping function in the OpenCV image editing library to remove the background, the shooting table and the surrounding environment objects;

[0019] Step S202: prepare a standard color card, place it together with the automobile plastic parts in the shooting scene for image acquisition, extract the color information of the standard color card area from the acquired image and express it in RGB color values, compare the extracted color card color values ​​with the known real color values ​​of the standard color card, calculate the color correction matrix, and use the calculated color correction matrix to correct the color of the entire image so that the image color is closer to the real color. The calculation formula is shown in formula (1):

[0020] Color = A × c (1)

[0021] Among them, Color is the corrected color value, A is the color correction matrix, and c is the original color value;

[0022] Step S203: performing image enhancement by mean filtering, including:

[0023] Step S2031: Determine the size of the filter kernel according to the noise level and detail features of the surface image of the automobile plastic part, process the three color channels R, G and B of RGB respectively from the upper left corner of the image, and traverse each pixel of the image according to the determined filter kernel size;

[0024] Step S2032: at each filter kernel position, read the pixel values ​​of all pixels in the filter kernel coverage area in the corresponding channel and add them up, and divide them by the total number of pixels in the filter kernel to obtain the mean pixel value of the channel;

[0025] Step S2033: The calculated pixel value mean is used as the new pixel value of the center pixel of the filter kernel in the channel and updated;

[0026] Step S2034: After completing the pixel value update of the current filter kernel position, move the filter kernel one pixel to the right, repeat steps S2032 and S2033, and when reaching the image boundary, wrap to the first pixel of the next row until the entire image is traversed to obtain the enhanced surface image of the automobile plastic part.

[0027] Preferably, the step of extracting the color features of the surface image data of the automobile plastic part in step S30 includes:

[0028] Color space conversion: The surface image of the automobile plastic part is converted from the RGB color space to the HSV color space through the cvtColor function in the OpenCV library. H represents hue, S represents saturation, and V represents brightness. After conversion, it is divided into three independent channels: H, S, and V.

[0029] Calculate the mean feature: Calculate the mean of the three channels H, S and V respectively. In the process of traversing the pixels in the plastic part area, accumulate the values ​​of the H, S and V channels of each pixel. The mean of the H channel is shown in formula (2), the mean of the S channel is shown in formula (3), and the mean of the V channel is shown in formula (4):

[0030]

[0031] in and are the means of the H, S and V channels respectively, ∑H, ∑S and ∑S are the accumulated values ​​of the H, S and V channels respectively, n is the total number of pixels in the surface image of the automobile plastic part, and the total number of pixels in each of the H, S and V channels is the same as the total number of pixels in the surface image of the automobile plastic part;

[0032] Calculate variance features: Calculate variance for the three channels H, S and V respectively. The variance of the H channel is shown in formula (5), the variance of the S channel is shown in formula (6), and the variance of the V channel is shown in formula (7):

[0033]

[0034]

[0035] where σ 2 H , σ 2 S and σ 2 V are the variances of the H, S and V channels respectively, H i , S i and V i are the H channel value, S channel value and V channel value of the i-th pixel respectively;

[0036] Histogram feature: Create an array Q of size 360 ​​for the H channel H , S channel and V channel create an array Q of size 256 S and Q V , the initial values ​​are all set to 0; traverse the pixel points in the image area of ​​the surface of the automobile plastic part, and the H channel value H of the i-th pixel iAs an index, the array Q H Add 1 to get the number of times the channel value appears in the H channel; for the S and V channels, add the S channel value S of the i-th pixel i and V channel value V i As an index, the array Q S and Q V Count the number of times the channel value appears in the S channel and the V channel; after traversing the pixel points in the image area of ​​the surface of the automobile plastic part, the histogram characteristics of the number of times each channel value appears in the H, S and V channels are obtained.

[0037] Among them, the mean feature, variance feature and histogram feature in the color feature are used to judge the degree of thermal oxygen aging of automobile plastic parts. The judgment principle is that with the thermal oxygen aging of automobile plastic parts, the color hue of the plastic parts shifts, and the saturation and brightness change. The changes in these color features can be used as an important indicator of the degree of aging. For example, after aging, white plastic parts turn yellow, and their hue value will change to the hue interval corresponding to yellow. At the same time, the saturation and brightness will also change accordingly, resulting in changes in the mean feature, variance feature and histogram in the color feature.

[0038] Preferably, the feature extraction of the shape features of the automobile plastic part surface image data in step S30 includes calculating the perimeter of the edge contour using the arcLength function in the OpenCV library based on the boundary coordinates and edge contour obtained in step S201, and calculating the area enclosed by the edge contour using the contourArea function in the OpenCV library.

[0039] Preferably, the step of constructing and training a thermal oxidation aging detection model for automobile plastic parts based on an improved convolutional neural network in step S40 includes:

[0040] Dataset construction: Obtain surface image samples of automobile plastic parts with different degrees of thermal oxidation aging, and divide them into training set, validation set and test set in a ratio of 80%:10%:10%;

[0041] Model construction: The entire network model includes an input layer, a convolution layer, a batch normalization layer, a maximum pooling layer, a residual block, and a fully connected layer; when the model is trained, the input of the input layer is the surface image samples of automobile plastic parts with different degrees of thermal oxidation aging. When the model is trained and applied, the input of the input layer is the surface image of automobile plastic parts after data preprocessing and feature extraction in step S30; the neuron connections between the fully connected layers are discarded at a ratio of 0.45 by the Dropout method;

[0042] Model training and verification: After the model is built, set the model parameters, use the divided training set as input to train the model, and use the divided verification set to verify the trained model after each training round;

[0043] Model evaluation: When overfitting occurs, stop model training and validation operations, and use the divided test set to evaluate the model;

[0044] Model optimization: Adjust model parameters and retrain according to model evaluation results until the optimal model parameter combination is obtained and the corresponding automotive plastic parts thermal oxygen aging detection model version is determined.

[0045] Preferably, after the model training in step S50 is completed, the thermal oxygen aging test of automobile plastic parts is carried out and the test results are output, the test results and the test time are stored. In addition, the remaining service life of the automobile plastic parts is estimated based on the test results, and a test report is generated. The test report includes the degree of thermal oxygen aging of automobile plastic parts, color feature change data, shape feature change data and remaining service life estimation information.

[0046] In addition, to achieve the above-mentioned purpose, the present invention also proposes a thermal oxygen aging detection system for automobile plastic parts, the thermal oxygen aging detection system for automobile plastic parts comprising:

[0047] Automobile plastic parts surface image shooting module: Use industrial cameras to shoot the surface of automobile plastic parts to obtain the surface image data of automobile plastic parts;

[0048] Automobile plastic parts surface image preprocessing module: performs data preprocessing on the obtained automobile plastic parts surface image data, including cropping, color correction and image enhancement;

[0049] Automobile plastic parts surface image feature extraction module: After data preprocessing, the color features and shape features of automobile plastic parts surface image data are extracted;

[0050] Automobile plastic parts thermal oxygen aging detection model construction and training module: After feature extraction is completed, the automobile plastic parts thermal oxygen aging detection model based on the improved convolutional neural network is constructed and trained;

[0051] Automobile plastic parts thermal oxygen aging detection module: After the model training is completed, the automobile plastic parts thermal oxygen aging detection is carried out and the detection results are output;

[0052] The industrial camera used in the automobile plastic part surface image shooting module has a 2k resolution, and a ring light source is installed above the automobile plastic part to keep the lighting uniform and stable during the shooting process;

[0053] The data preprocessing steps in the automobile plastic part surface image preprocessing module include cropping, color correction and image enhancement. Cropping is used to remove the background, the shooting table and the surrounding objects in the automobile plastic part surface image. Color correction is used to restore the true value of the image data. Image enhancement is used to remove noise in the image to make the image features more obvious.

[0054] The color extraction in the automobile plastic part surface image feature extraction module is HSV color feature extraction, including hue, saturation and brightness; the shape feature extraction is performed by obtaining the edge contour of the plastic part from the cropped automobile plastic part surface image and calculating the perimeter and area of ​​the edge contour;

[0055] The improved convolutional neural network in the automobile plastic parts thermal oxidation aging detection model construction and training module adopts a multi-level convolutional network architecture, 64 3×3 convolution kernels are added to the first convolution layer to extract local detail features of the automobile plastic parts surface image, and a batch normalization layer is introduced after each convolution layer to standardize the data; in addition, residual connection is adopted in the network structure to construct a residual block, each residual block contains 2 convolution layers, the convolution kernel sizes are 3×3 and 1×1, and the number of convolution kernels is 128 and 64, respectively. A batch normalization layer is added between the two convolution layers, and ReLU is used as the activation function; the full connection layer is set to 3 layers, and the number of neurons in each layer is 256, 128 and 64, respectively. The neuron connections are discarded according to the set ratio between the full connection layers through the Dropout method to prevent overfitting, and the last layer outputs the classification results of the thermal oxidation aging of automobile plastic parts. The Softmax function is used for multi-classification output, and the output is converted into the probability distribution of each category.

[0056] In addition, to achieve the above-mentioned purpose, the present invention also proposes a thermal oxygen aging detection device for automotive plastic parts, the device comprising: a memory, a processor, and programs such as an automotive plastic parts thermal oxygen aging detection algorithm based on an improved convolutional neural network stored in the memory and run on the processor, the programs such as the automotive plastic parts thermal oxygen aging detection algorithm based on an improved convolutional neural network are steps to implement a method for thermal oxygen aging detection of automotive plastic parts as described above.

[0057] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer program product, which includes programs such as an automotive plastic parts thermal oxygen aging detection algorithm based on an improved convolutional neural network. When the programs such as the automotive plastic parts thermal oxygen aging detection algorithm based on an improved convolutional neural network are executed by a processor, a method for detecting thermal oxygen aging of automotive plastic parts as described above is implemented.

[0058] The advantages and effects of the present invention are:

[0059] The present invention proposes a method for detecting thermal oxygen aging of automobile plastic parts. The method adopts image data processing technology combined with an improved convolutional neural network method to accurately detect the thermal oxygen aging conditions of automobile plastic parts by capturing multiple features. Through automated image processing, no human intervention is required, which greatly saves time and labor costs, shortens the detection time of thermal oxygen aging of automobile plastic parts, and improves detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0061] Figure 1 The present invention is a flow chart of a method for detecting thermal oxidation aging of automobile plastic parts.

[0062] Figure 2 The present invention is a schematic structural diagram of a thermal oxygen aging detection system for automobile plastic parts.

[0063] Figure 3 The present invention is a schematic block diagram of the structure of an electronic device for detecting thermal oxygen aging of automobile plastic parts. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] The present invention provides a method for detecting thermal oxidation aging of automobile plastic parts. Figure 1 As shown, the following steps are included:

[0066] Step S10: The surface of the automobile plastic part is photographed by an industrial camera with high resolution and high color reproduction to obtain surface image data of the automobile plastic part. During the acquisition process, uniform and stable lighting is ensured to avoid the influence of shadows and reflections on the image quality.

[0067] The industrial camera used in step S10 has a 2k resolution, and a ring light source is installed above the automobile plastic parts to maintain uniform and stable lighting during the shooting process.

[0068] Step S20: performing data preprocessing on the obtained automobile plastic part surface image data, including cropping, color correction and image enhancement.

[0069] The data preprocessing steps in step S20 include cropping, color correction and image enhancement. Cropping is used to remove the background, shooting table and surrounding objects in the surface image of the automobile plastic part to reduce the amount of data processing; color correction is used to restore the true value of the image data, and the correction formula can be calibrated according to the standard color card; image enhancement is used to remove noise in the surface image of the automobile plastic part and enhance the recognition of aging features in the image.

[0070] Specifically, in step S20, data preprocessing is performed on the obtained automobile plastic part surface image data, including:

[0071] Step S201: According to the position and outline of the automobile plastic part in the image, the boundary coordinates are determined, the shape of the edge outline of the automobile plastic part is obtained by the Canny edge detection algorithm, and the cropping area is set according to the obtained boundary coordinates and edge outline, and the surface image of the automobile plastic part is cropped by using the cropping function in the OpenCV image editing library to remove the background, the shooting table and the surrounding environment objects;

[0072] Step S202: prepare a standard color card, place it together with the automobile plastic parts in the shooting scene for image acquisition, extract the color information of the standard color card area from the acquired image and express it in RGB color values, compare the extracted color card color values ​​with the known real color values ​​of the standard color card, calculate the color correction matrix, and use the calculated color correction matrix to correct the color of the entire image so that the image color is closer to the real color. The calculation formula is shown in formula (1):

[0073] Color = A × c (1)

[0074] Among them, Color is the corrected color value, A is the color correction matrix, and c is the original color value;

[0075] Step S203: performing image enhancement by mean filtering, including:

[0076] Step S2031: Determine the size of the filter kernel according to the noise level and detail features of the surface image of the automobile plastic part, process the three color channels R, G and B of RGB respectively from the upper left corner of the image, and traverse each pixel of the image according to the determined filter kernel size;

[0077] Step S2032: at each filter kernel position, read the pixel values ​​of all pixels in the filter kernel coverage area in the corresponding channel and add them up, and divide them by the total number of pixels in the filter kernel to obtain the mean pixel value of the channel;

[0078] Step S2033: The calculated pixel value mean is used as the new pixel value of the center pixel of the filter kernel in the channel and updated;

[0079] Step S2034: After completing the pixel value update of the current filter kernel position, move the filter kernel one pixel to the right, repeat steps S2032 and S2033, and when reaching the image boundary, wrap to the first pixel of the next row until the entire image is traversed to obtain the enhanced surface image of the automobile plastic part.

[0080] Step S30: After the data preprocessing is completed, the color features and shape features of the automobile plastic part surface image data are extracted.

[0081] In step S30, the color extraction is HSV color feature extraction, including hue, saturation and brightness; in the shape feature extraction, the edge contour of the plastic part is obtained by the cropped surface image of the automobile plastic part and the perimeter and area of ​​the edge contour are calculated.

[0082] Specifically, the step of extracting the color features of the surface image data of the automobile plastic part in step S30 includes:

[0083] Color space conversion: The surface image of the automobile plastic part is converted from the RGB color space to the HSV color space through the cvtColor function in the OpenCV library. H represents hue, S represents saturation, and V represents brightness. After conversion, it is divided into three independent channels: H, S, and V.

[0084] Calculate the mean feature: Calculate the mean of the three channels H, S and V respectively. In the process of traversing the pixels in the plastic part area, accumulate the values ​​of the H, S and V channels of each pixel. The mean of the H channel is shown in formula (2), the mean of the S channel is shown in formula (3), and the mean of the V channel is shown in formula (4):

[0085]

[0086] in and are the means of the H, S and V channels respectively, ∑H, ∑S and ∑S are the accumulated values ​​of the H, S and V channels respectively, n is the total number of pixels in the surface image of the automobile plastic part, and the total number of pixels in each of the H, S and V channels is the same as the total number of pixels in the surface image of the automobile plastic part;

[0087] Calculate variance features: Calculate variance for the three channels H, S and V respectively. The variance of the H channel is shown in formula (5), the variance of the S channel is shown in formula (6), and the variance of the V channel is shown in formula (7):

[0088]

[0089] where σ 2 H , σ 2S and σ 2 V are the variances of the H, S and V channels respectively, H i , S i and V i are the H channel value, S channel value and V channel value of the i-th pixel respectively;

[0090] Histogram feature: Create an array Q of size 360 ​​for the H channel H , S channel and V channel create an array Q of size 256 S and Q V , the initial values ​​are all set to 0; traverse the pixel points in the image area of ​​the surface of the automobile plastic part, and the H channel value H of the i-th pixel i As an index, the array Q H Add 1 to get the number of times the channel value appears in the H channel; for the S and V channels, add the S channel value S of the i-th pixel i and V channel value V i As an index, the array Q S and Q V Count the number of times the channel value appears in the S channel and the V channel; after traversing the pixel points in the image area of ​​the surface of the automobile plastic part, the histogram characteristics of the number of times each channel value appears in the H, S and V channels are obtained.

[0091] Among them, the mean feature, variance feature and histogram feature in the color feature are used to judge the degree of thermal oxygen aging of automobile plastic parts. The judgment principle is that with the thermal oxygen aging of automobile plastic parts, the color hue of the plastic parts shifts, and the saturation and brightness change. The changes in these color features can be used as an important indicator of the degree of aging. For example, after aging, white plastic parts turn yellow, and their hue value will change to the hue interval corresponding to yellow. At the same time, the saturation and brightness will also change accordingly, resulting in changes in the mean feature, variance feature and histogram in the color feature.

[0092] Specifically, the feature extraction of the shape features of the automobile plastic part surface image data in step S30 includes calculating the perimeter of the edge contour using the arcLength function in the OpenCV library according to the boundary coordinates and edge contour obtained in step S201, and calculating the area enclosed by the edge contour using the contourArea function in the OpenCV library.

[0093] Step S40: After feature extraction is completed, a thermal oxidation aging detection model for automotive plastic parts based on an improved convolutional neural network is constructed and trained.

[0094] In step S40, the improved convolutional neural network adopts a multi-level convolutional network architecture, and 64 3×3 convolution kernels are added to the first convolution layer to extract local detail features of the surface image of automobile plastic parts, and a batch normalization layer is introduced after each convolution layer to standardize the data; in addition, residual connection is used in the network structure to construct a residual block, each residual block contains 2 convolution layers, the convolution kernel sizes are 3×3 and 1×1, and the number of convolution kernels is 128 and 64, respectively. A batch normalization layer is added between the two convolution layers, and ReLU is used as the activation function; the fully connected layer is set to 3 layers, and the number of neurons in each layer is 256, 128 and 64, respectively. The Dropout method is used between the fully connected layers to discard neuronal connections according to the set ratio to prevent overfitting. The output of the last layer is the classification result of thermal oxidation aging of automobile plastic parts. The Softmax function is used for multi-classification output, and the output is converted into the probability distribution of each category.

[0095] Specifically, the steps of constructing and training the automotive plastic parts thermal oxidation aging detection model based on the improved convolutional neural network in step S40 include:

[0096] Dataset construction: Obtain surface image samples of automobile plastic parts with different degrees of thermal oxidation aging, and divide them into training set, validation set and test set in a ratio of 80%:10%:10%;

[0097] Model construction: The entire network model includes an input layer, a convolution layer, a batch normalization layer, a maximum pooling layer, a residual block, and a fully connected layer; when the model is trained, the input of the input layer is the surface image samples of automobile plastic parts with different degrees of thermal oxidation aging. When the model is trained and applied, the input of the input layer is the surface image of automobile plastic parts after data preprocessing and feature extraction in step S30; the neuron connections between the fully connected layers are discarded at a ratio of 0.45 by the Dropout method;

[0098] Model training and verification: After the model is built, set the model parameters, use the divided training set as input to train the model, and use the divided verification set to verify the trained model after each training round;

[0099] Model evaluation: When overfitting occurs, stop model training and validation operations, and use the divided test set to evaluate the model;

[0100] Model optimization: Adjust model parameters and retrain according to model evaluation results until the optimal model parameter combination is obtained and the corresponding automotive plastic parts thermal oxygen aging detection model version is determined.

[0101] Step S50: After the model training is completed, thermal oxygen aging test of automobile plastic parts is performed and the test results are output.

[0102] Specifically, after the model training is completed in step S50, the thermal oxygen aging test of automobile plastic parts is carried out and the test results are output, the test results and the test time are stored. In addition, the remaining service life of the automobile plastic parts is estimated based on the test results, and a test report is generated. The test report includes the degree of thermal oxygen aging of automobile plastic parts, color feature change data, shape feature change data and remaining service life estimation information.

[0103] In addition, the present invention also proposes a thermal oxygen aging detection system for automotive plastic parts, please refer to Figure 2 The automotive plastic parts thermal oxygen aging detection system comprises:

[0104] Automobile plastic parts surface image shooting module: Use industrial cameras to shoot the surface of automobile plastic parts to obtain the surface image data of automobile plastic parts;

[0105] Automobile plastic parts surface image preprocessing module: performs data preprocessing on the obtained automobile plastic parts surface image data, including cropping, color correction and image enhancement;

[0106] Automobile plastic parts surface image feature extraction module: After data preprocessing, the color features and shape features of automobile plastic parts surface image data are extracted;

[0107] Automobile plastic parts thermal oxygen aging detection model construction and training module: After feature extraction is completed, the automobile plastic parts thermal oxygen aging detection model based on the improved convolutional neural network is constructed and trained;

[0108] Automobile plastic parts thermal oxygen aging detection module: After the model training is completed, the automobile plastic parts thermal oxygen aging detection is carried out and the detection results are output;

[0109] Among them, the industrial camera used in the automobile plastic parts surface image shooting module has a 2k resolution, and a ring light source is installed above the automobile plastic parts to keep the lighting uniform and stable during the shooting process;

[0110] Among them, the data preprocessing steps in the automobile plastic parts surface image preprocessing module include cropping, color correction and image enhancement. Cropping is used to remove the background, shooting table and surrounding objects in the automobile plastic parts surface image. Color correction is used to restore the true value of the image data. Image enhancement is used to remove noise in the image and make the image features more obvious.

[0111] Among them, the color extraction in the automobile plastic part surface image feature extraction module is HSV color feature extraction, including hue, saturation and brightness; in the shape feature extraction, the edge contour of the plastic part is obtained by the cropped automobile plastic part surface image and the perimeter and area of ​​the edge contour are calculated;

[0112] Among them, the improved convolutional neural network in the construction and training module of the thermal oxidation aging detection model for automotive plastic parts adopts a multi-level convolutional network architecture. 64 3×3 convolution kernels are added to the first convolution layer to extract local detail features of the surface image of automotive plastic parts, and a batch normalization layer is introduced after each convolution layer to standardize the data; in addition, residual connection is used in the network structure to construct residual blocks. Each residual block contains 2 convolution layers, and the convolution kernel sizes are 3×3 and 1×1, respectively, and the number of convolution kernels is 128 and 64, respectively. A batch normalization layer is added between the two convolution layers, and ReLU is used as the activation function; the fully connected layer is set to 3 layers, and the number of neurons in each layer is 256, 128 and 64, respectively. The Dropout method is used between the fully connected layers to discard neuronal connections according to the set ratio to prevent overfitting. The output of the last layer is the classification result of thermal oxidation aging of automotive plastic parts. The Softmax function is used for multi-classification output, and the output is converted into the probability distribution of each category.

[0113] The present application provides a thermal oxygen aging detection system for automotive plastic parts, which adopts a thermal oxygen aging detection method for automotive plastic parts in the above embodiment, and can solve the technical problems of low efficiency and low accuracy of traditional thermal oxygen aging detection methods for automotive plastic parts. Compared with the prior art, the beneficial effects of the thermal oxygen aging detection system for automotive plastic parts provided by the present application are the same as the beneficial effects of the thermal oxygen aging detection method for automotive plastic parts provided by the above embodiment, and other technical features of the thermal oxygen aging detection system for automotive plastic parts are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0114] The present application provides a thermal oxygen aging detection device for automobile plastic parts, the thermal oxygen aging detection device for automobile plastic parts comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a thermal oxygen aging detection method for automobile plastic parts in the above-mentioned embodiment 1.

[0115] Reference below Figure 3, which shows a schematic diagram of the structure of a thermal oxygen aging detection device for automotive plastic parts suitable for implementing the embodiment of the present application. The thermal oxygen aging detection device for automotive plastic parts in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The automotive plastic parts thermal oxygen aging detection equipment shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0116] Figure 3 The thermal oxygen aging detection device for automobile plastic parts shown may include a processing system 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage system 1003 to a random access memory (RAM) 1004. In RAM 1004, various programs and data required for the operation of a thermal oxygen aging detection device for automobile plastic parts are also stored. The processing system 1001, ROM 1002, and RAM 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: an input system 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output system 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage system 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication system 1009. The communication system 1009 can allow a thermal oxygen aging detection device for automotive plastic parts to communicate wirelessly or wired with other devices to exchange data. Although a thermal oxygen aging detection device for automotive plastic parts with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0117] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication system, or installed from a storage system 1003, or installed from a ROM 1002. When the computer program is executed by the processing system 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0118] The present application provides a thermal oxygen aging detection device for automotive plastic parts, which adopts a thermal oxygen aging detection method for automotive plastic parts in the above embodiment, and can solve the technical problems of low efficiency and low accuracy of traditional thermal oxygen aging detection methods for automotive plastic parts. Compared with the prior art, the beneficial effects of the thermal oxygen aging detection device for automotive plastic parts provided by the present application are the same as the beneficial effects of the thermal oxygen aging detection method for automotive plastic parts provided by the above embodiment, and the other technical features of the thermal oxygen aging detection device for automotive plastic parts are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0119] The various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0120] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned method for detecting thermal oxidation aging of automobile plastic parts when executed by a processor.

[0121] The computer program product provided by the present application can solve the technical problems of low efficiency and low accuracy of the traditional thermal oxygen aging detection method for automobile plastic parts. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the thermal oxygen aging detection method for automobile plastic parts provided by the above embodiment, which will not be described in detail here.

[0122] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for detecting thermal oxidation aging of automobile plastic parts, characterized in that: The method comprises the following steps: Step S10: photographing the surface of the automobile plastic part by an industrial camera to obtain surface image data of the automobile plastic part; Step S20: performing data preprocessing on the obtained automobile plastic part surface image data, including cropping, color correction and image enhancement; Step S30: after the data preprocessing is completed, feature extraction is performed on the color features and shape features of the surface image data of the automobile plastic part; Step S40: after feature extraction is completed, a thermal oxidation aging detection model for automobile plastic parts based on an improved convolutional neural network is constructed and trained; Step S50: After the model training is completed, thermal oxygen aging test of automobile plastic parts is performed and the test results are output; The industrial camera used in step S10 has a 2k resolution, and a ring light source is installed above the automobile plastic part to keep the lighting uniform and stable during the shooting process; The data preprocessing step in step S20 includes cropping, color correction and image enhancement, wherein cropping is used to remove the background, the shooting table and the surrounding objects in the surface image of the automobile plastic part, color correction is used to restore the true value of the image data, and image enhancement is used to remove noise in the image; In the step S30, the color extraction is HSV color feature extraction, including hue, saturation and brightness; in the shape feature extraction, the edge contour of the plastic part is obtained by the cropped surface image of the automobile plastic part and the perimeter and area of ​​the edge contour are calculated; The improved convolutional neural network in step S40 adopts a multi-level convolutional network architecture, 64 3×3 convolution kernels are added to the first convolution layer to extract local detail features of the surface image of automobile plastic parts, and a batch normalization layer is introduced after each convolution layer to standardize the data; in addition, residual connection is used in the network structure to construct a residual block, each residual block contains 2 convolution layers, the convolution kernel sizes are 3×3 and 1×1, and the number of convolution kernels is 128 and 64, respectively. A batch normalization layer is added between the two convolution layers, and ReLU is used as the activation function; the fully connected layer is set to 3 layers, and the number of neurons in each layer is 256, 128 and 64, respectively. The neuron connections are discarded according to the set ratio between the fully connected layers through the Dropout method, and the last layer outputs the classification results of thermal oxidation aging of automobile plastic parts. The Softmax function is used for multi-classification output, and the output is converted into the probability distribution of each category.

2. The method for detecting thermal oxidation aging of automobile plastic parts according to claim 1, characterized in that: The step S20 performs data preprocessing on the obtained automobile plastic part surface image data, including: Step S201: According to the position and outline of the automobile plastic part in the image, the boundary coordinates are determined, the shape of the edge outline of the automobile plastic part is obtained by the Canny edge detection algorithm, and the cropping area is set according to the obtained boundary coordinates and edge outline, and the surface image of the automobile plastic part is cropped by using the cropping function in the OpenCV image editing library to remove the background, the shooting table and the surrounding environment objects; Step S202: prepare a standard color card, place it together with the automobile plastic parts in the shooting scene for image acquisition, extract the color information of the standard color card area from the acquired image and express it in RGB color values, compare the extracted color card color values ​​with the known true color values ​​of the standard color card, calculate the color correction matrix, and use the calculated color correction matrix to correct the color of the entire image. The calculation formula is shown in formula (1): Color = A × c (1) Among them, Color is the corrected color value, A is the color correction matrix, and c is the original color value; Step S203: performing image enhancement by mean filtering, including: Step S2031: Determine the size of the filter kernel according to the noise level and detail features of the surface image of the automobile plastic part, process the three color channels R, G and B of RGB respectively from the upper left corner of the image, and traverse each pixel of the image according to the determined filter kernel size; Step S2032: at each filter kernel position, read the pixel values ​​of all pixels in the filter kernel coverage area in the corresponding channel and add them up, and divide them by the total number of pixels in the filter kernel to obtain the mean pixel value of the channel; Step S2033: The calculated pixel value mean is used as the new pixel value of the center pixel of the filter kernel in the channel and updated; Step S2034: After completing the pixel value update of the current filter kernel position, move the filter kernel one pixel to the right, repeat steps S2032 and S2033, and when reaching the image boundary, wrap to the first pixel of the next row until the entire image is traversed to obtain the enhanced surface image of the automobile plastic part.

3. The method for detecting thermal oxidation aging of automobile plastic parts according to claim 1, characterized in that: The step of extracting the color features of the surface image data of the automobile plastic part in step S30 includes: Color space conversion: The surface image of the automobile plastic part is converted from the RGB color space to the HSV color space through the cvtColor function in the OpenCV library. H represents hue, S represents saturation, and V represents brightness. After conversion, it is divided into three independent channels: H, S, and V. Calculate the mean feature: Calculate the mean of the three channels H, S and V respectively. In the process of traversing the pixels in the plastic part area, accumulate the values ​​of the H, S and V channels of each pixel. The mean of the H channel is shown in formula (2), the mean of the S channel is shown in formula (3), and the mean of the V channel is shown in formula (4): in and are the means of the H, S and V channels respectively, ∑H, ∑S and ∑S are the accumulated values ​​of the H, S and V channels respectively, n is the total number of pixels in the surface image of the automobile plastic part, and the total number of pixels in each of the H, S and V channels is the same as the total number of pixels in the surface image of the automobile plastic part; Calculate variance features: Calculate variance for the three channels H, S and V respectively. The variance of the H channel is shown in formula (5), the variance of the S channel is shown in formula (6), and the variance of the V channel is shown in formula (7): where σ 2 H , σ 2 S and σ 2 V are the variances of the H, S and V channels respectively, H i , S i and V i are the H channel value, S channel value and V channel value of the i-th pixel respectively; Histogram feature: Create an array Q of size 360 ​​for the H channel H , S channel and V channel create an array Q of size 256 S and Q V , the initial values ​​are all set to 0; traverse the pixel points in the image area of ​​the surface of the automobile plastic part, and the H channel value H of the i-th pixel i As an index, the array Q H Add 1 to get the number of times the channel value appears in the H channel; for the S and V channels, add the S channel value S of the i-th pixel i and V channel value V i As an index, the array Q S and Q V Count the number of times the channel value appears in the S channel and the V channel; after traversing the pixel points in the image area of ​​the surface of the automobile plastic part, the histogram characteristics of the number of times each channel value appears in the H, S and V channels are obtained.

4. The method for detecting thermal oxidation aging of automobile plastic parts according to claim 3, characterized in that: The mean feature, variance feature and histogram feature in the color feature are used to judge the degree of thermal oxidation aging of automobile plastic parts.

5. The method for detecting thermal oxidation aging of automobile plastic parts according to claim 1, characterized in that: The feature extraction of the shape features of the automobile plastic part surface image data in step S30 includes calculating the perimeter of the edge contour using the arcLength function in the OpenCV library according to the boundary coordinates and edge contour obtained in step S201, and calculating the area enclosed by the edge contour using the contourArea function in the OpenCV library.

6. The method for detecting thermal oxidation aging of automobile plastic parts according to claim 1, characterized in that: The step of constructing and training a thermal oxidation aging detection model for automobile plastic parts based on an improved convolutional neural network in step S40 includes: Dataset construction: Obtain surface image samples of automobile plastic parts with different degrees of thermal oxidation aging, and divide them into training set, validation set and test set in a ratio of 80%:10%:10%; Model construction: The entire network model includes an input layer, a convolution layer, a batch normalization layer, a maximum pooling layer, a residual block, and a fully connected layer; when the model is trained, the input of the input layer is the surface image samples of automobile plastic parts with different degrees of thermal oxidation aging. When the model is trained and applied, the input of the input layer is the surface image of automobile plastic parts after data preprocessing and feature extraction in step S30; the neuron connections between the fully connected layers are discarded at a ratio of 0.45 by the Dropout method; Model training and verification: After the model is built, set the model parameters, use the divided training set as input to train the model, and use the divided verification set to verify the trained model after each training round; Model evaluation: When overfitting occurs, stop model training and validation operations, and use the divided test set to evaluate the model; Model optimization: Adjust model parameters and retrain according to model evaluation results until the optimal model parameter combination is obtained and the corresponding automotive plastic parts thermal oxygen aging detection model version is determined.

7. The method for detecting thermal oxidation aging of automobile plastic parts according to claim 1, characterized in that: After the model training is completed in step S50, the thermal oxygen aging test of automobile plastic parts is performed and the test results are output, the test results and the test time are stored, and the remaining service life of the automobile plastic parts is estimated based on the test results, and a test report is generated. The test report includes the degree of thermal oxygen aging of automobile plastic parts, color feature change data, shape feature change data and remaining service life estimation information.

8. A thermal oxygen aging detection system for automobile plastic parts, characterized in that: The automotive plastic parts thermal oxygen aging detection system comprises: Automobile plastic parts surface image shooting module: Use industrial cameras to shoot the surface of automobile plastic parts to obtain the surface image data of automobile plastic parts; Automobile plastic parts surface image preprocessing module: performs data preprocessing on the obtained automobile plastic parts surface image data, including cropping, color correction and image enhancement; Automobile plastic parts surface image feature extraction module: After data preprocessing, the color features and shape features of automobile plastic parts surface image data are extracted; Automobile plastic parts thermal oxygen aging detection model construction and training module: After feature extraction is completed, the automobile plastic parts thermal oxygen aging detection model based on the improved convolutional neural network is constructed and trained; Automobile plastic parts thermal oxygen aging detection module: After the model training is completed, the automobile plastic parts thermal oxygen aging detection is carried out and the detection results are output; The industrial camera used in the automobile plastic part surface image shooting module has a 2k resolution, and a ring light source is installed above the automobile plastic part to keep the lighting uniform and stable during the shooting process; The data preprocessing steps in the automobile plastic part surface image preprocessing module include cropping, color correction and image enhancement, wherein cropping is used to remove the background, the shooting table and the surrounding objects in the automobile plastic part surface image, color correction is used to restore the true value of the image data, and image enhancement is used to remove noise in the image; The color extraction in the automobile plastic part surface image feature extraction module is HSV color feature extraction, including hue, saturation and brightness; the shape feature extraction is performed by obtaining the edge contour of the plastic part from the cropped automobile plastic part surface image and calculating the perimeter and area of ​​the edge contour; The improved convolutional neural network in the automobile plastic parts thermal oxidation aging detection model construction and training module adopts a multi-level convolutional network architecture, 64 3×3 convolution kernels are added to the first convolution layer to extract local detail features of the automobile plastic parts surface image, and a batch normalization layer is introduced after each convolution layer to standardize the data; in addition, residual connection is adopted in the network structure to construct a residual block, each residual block contains 2 convolution layers, the convolution kernel sizes are 3×3 and 1×1, and the number of convolution kernels is 128 and 64, respectively. A batch normalization layer is added between the two convolution layers, and ReLU is used as the activation function; the full connection layer is set to 3 layers, and the number of neurons in each layer is 256, 128 and 64, respectively. The neuron connections are discarded according to the set ratio between the full connection layers through the Dropout method to prevent overfitting, and the last layer outputs the classification results of the thermal oxidation aging of automobile plastic parts. The Softmax function is used for multi-classification output, and the output is converted into the probability distribution of each category.

9. A thermal oxygen aging testing device for automobile plastic parts, characterized in that: The automotive plastic parts thermal oxygen aging detection equipment comprises: A memory, a processor, and a thermal oxygen aging detection program for automobile plastic parts stored in the memory and executable on the processor, wherein the thermal oxygen aging detection program for automobile plastic parts, when executed by the processor, implements a thermal oxygen aging detection method for automobile plastic parts as described in any one of claims 1 to 7.

10. A computer program product, characterized in that The computer program product comprises a thermal oxygen aging detection program for automobile plastic parts. When the thermal oxygen aging detection program for automobile plastic parts is executed by a processor, an automobile plastic part thermal oxygen aging detection method according to any one of claims 1 to 8 is implemented.

Images