Bubble imaging detection system for interior of brake pad
By using convolutional neural network to classify bubbles in the internal imaging detection system of the brake pad, the problem of subjectivity and low efficiency of traditional detection methods is solved, and efficient and accurate detection and classification of the internal bubbles of the brake pad is achieved, ensuring the vehicle Safety.
Patent Information
- Application Number
- CN202410943787.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-15
- Publication Date
- 2024-11-08
AI Technical Summary
The traditional knock sound detection method has strong subjectivity and low efficiency in the detection of internal bubbles on the brake pads. It cannot effectively classify different types of bubbles, making it difficult to meet the needs of modern automobile manufacturing for efficient and accurate detection.
Design a bubble imaging detection system to obtain images through the internal imaging module of the brake pad, use the bubble recognition module to enhance and segment, and classify bubbles through convolutional neural networks to build a bubble distribution density evaluation model to realize automatic bubble recognition and Classification.
It improves the accuracy and efficiency of detection, can promptly detect potential defects, ensure the quality of brake pads and vehicle safety, and achieves accurate classification of different types of bubbles.
Smart Images

Figure CN118914193A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field, and more particularly to a bubble imaging detection system for the interior of a brake pad. Background Art
[0002] In the process of automobile manufacturing and maintenance, brake pads are one of the key components to ensure the safe operation of vehicles. The performance of brake pads directly affects the braking effect and driving safety of automobiles. Brake pads can be divided into asbestos friction material brake pads, semi-metallic friction material brake pads, low-metal friction material brake pads, organic friction material brake pads and carbon fiber friction material brake pads according to the composition of friction materials; among them, low-metal friction material brake pads have reduced metal components, good performance in all aspects, high cost performance, and are widely used in cars and commercial vehicles and gradually become the mainstream. However, due to manufacturing process and other reasons, there may be bubble defects inside low-metal friction material brake pads. These internal bubbles will cause large-area friction materials to fall off, thereby affecting the braking effect and greatly threatening driving safety. Therefore, timely and accurate detection of bubbles inside low-metal friction material brake pads has become an important link to ensure the quality of brake pads and vehicle safety.
[0003] The traditional percussion and listening method is the main means of detecting bubbles inside the brake pad. This method relies on the operator's experience and subjective judgment, and judges whether there are bubbles inside by percussing the brake pad and listening to the changes in the sound. However, the detection results of this method are highly dependent on the operator's experience and hearing, and are highly subjective and have poor repeatability. Secondly, the percussion and listening method can only detect larger bubbles, and has poor detection effects on tiny bubbles and deep bubbles. Different types of bubbles have different effects on the brake pad, and the traditional percussion and listening method cannot classify bubbles. In addition, the detection efficiency is low, which makes it difficult to meet the needs of modern automobile manufacturing for efficient and accurate detection.
[0004] In order to solve the above problems, a technical solution is now provided. Summary of the invention
[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a bubble imaging detection system for the inside of a brake pad, which acquires images through an imaging module inside the brake pad, performs image enhancement and segmentation using a bubble recognition module, and classifies bubbles through a convolutional neural network to construct a bubble distribution density assessment model; the problems of traditional percussion and sound detection methods, such as strong subjectivity, low detection efficiency, and inability to classify different types of bubbles, are solved.
[0006] To achieve the above object, the present invention provides the following technical solutions:
[0007] A bubble imaging detection system for the inside of a brake pad comprises an imaging module inside the brake pad, a bubble recognition module, a bubble classification module and a user interface, wherein the imaging module inside the brake pad is connected to the bubble recognition module, the bubble recognition module is connected to the bubble classification module, and the bubble classification module is connected to the user interface; the bubble recognition module comprises a bubble image enhancement unit, a bubble image segmentation unit and a bubble parameter calculation unit, the bubble image enhancement unit is connected to the bubble image segmentation unit, the bubble image segmentation unit is connected to the bubble parameter calculation unit, and is used to identify and analyze image data acquired by the imaging module inside the brake pad, obtain bubble area parameters, and construct a bubble distribution density evaluation model according to the acquired parameters; the bubble recognition module is used to construct a bubble classification model according to a convolutional neural network, and use the image data of bubbles inside a metal-free friction material brake pad and the annotation information of the bubble area to identify the probability distribution of bubbles in corresponding category labels and obtain bubble classification results; wherein the formula of the bubble distribution density evaluation model is:
[0008]
[0009] Where D is the estimated value of the bubble distribution density, M is the number of pixels in the vertical direction of the image, N is the number of pixels in the horizontal direction of the image, and N is the number of pixels in the horizontal direction of the image. b is the total number of bubble regions in the image, A i is the area of the ith bubble, x i is the horizontal coordinate of the center position of the ith bubble, y i is the ordinate of the center position of the i-th bubble, A max is the maximum area among all bubbles.
[0010] As a further solution of the present invention, the bubble image enhancement unit is used to enhance the image acquired by the imaging module inside the brake pad and transmit it to the bubble image segmentation unit; the bubble image segmentation unit is used to segment the enhanced image, and when the pixel value of the enhanced image at the position (x, y) is greater than the threshold, the position is determined to be a bubble area and marked as 1; when the pixel value of the enhanced image at the position (x, y) is less than or equal to the threshold, the position is determined to be a non-bubble area and marked as 0; and a binary image is obtained.
[0011] As a further solution of the present invention, the bubble parameter calculation unit identifies the bubble area in the binary image through the image connected domain analysis technology, obtains the bubble area parameters of each bubble, including the area and perimeter of each bubble, and constructs a bubble relative number index evaluation model and a bubble morphological characteristic index evaluation model, wherein the formula of the bubble relative number index evaluation model is:
[0012]
[0013] Where Nb is the bubble relative number index evaluation value, M is the number of pixels of the image in the vertical direction, N is the number of pixels of the image in the horizontal direction, B(x,y) is a binary image, where B(x,y)=1 represents the bubble area, and B(x,y)=0 represents the non-bubble area;
[0014] The formula of the bubble morphology characteristic index evaluation model is:
[0015]
[0016] Where S is the evaluation value of the bubble morphological characteristic index, M is the number of pixels of the image in the vertical direction, N is the number of pixels of the image in the horizontal direction, B(x,y) is a binary image, where B(x,y)=1 represents the bubble area, B(x,y)=0 represents the non-bubble area, A(x,y) is the area of the bubble at position (x,y), and P(x,y) is the circumference of the bubble at position (x,y).
[0017] As a further scheme of the present invention, the bubble classification module is based on the enhanced image and the segmented bubble area, and inputs the constructed convolutional neural network to obtain the probability distribution of each bubble and the corresponding category label, and obtains the final bubble classification result by taking the category label corresponding to the maximum probability. The category labels include Class I bubbles, Class II bubbles, Class III bubbles, Class IV bubbles and Class V bubbles.
[0018] As a further solution of the present invention, the steps of constructing a convolutional neural network are:
[0019] Step S1, determining that the input layer includes image data of bubbles inside a low-metal friction material brake pad and annotation information of the bubble area, wherein the image data includes an original image, an enhanced image, and a segmented bubble area image, and the annotation information includes a relative quantity index evaluation value, a morphological feature index evaluation value, a distribution density evaluation value, a position coordinate, an area, and a perimeter of the bubble;
[0020] Step S2, cleaning and standardizing the above data to extract image features and bubble feature vectors;
[0021] Step S3, dividing the historical data into a training set and a validation set, using the training set to train a convolutional neural network algorithm model, and building a bubble classification model;
[0022] Step S4, after the training is completed, the bubble classification model is verified using the verification set;
[0023] Step S5, in the new bubble detection task, the enhanced image and the segmented bubble area image are input, the bubble classification model is used for prediction, and the corresponding bubble classification result is output.
[0024] As a further solution of the invention, the brake pad internal imaging module is used to obtain an image of the internal structure of the low-metal friction material brake pad, including: a light source for emitting a light beam into the interior of the low-metal friction material brake pad; an optical sensor for receiving the light beam reflected and transmitted through the interior of the low-metal friction material brake pad and converting it into an electrical signal; and an image processing unit for converting the electrical signal of the optical sensor into image data.
[0025] As a further solution of the present invention, the user interface is used to display the image data of bubbles inside the low-metal friction material brake pad and the annotation information of the bubble area output by the brake pad internal imaging module and the bubble recognition module, as well as the bubble classification results output by the bubble classification module.
[0026] Compared with the prior art, the bubble imaging detection system for the inside of the brake pad of the present invention has the following beneficial effects:
[0027] The present invention can obtain and analyze the bubble distribution inside the low-metal friction material brake pad by adopting the brake pad internal imaging module and the bubble recognition module; effectively identify the bubble area by using the image enhancement and segmentation technology, and comprehensively evaluate the number, area, shape and distribution of the bubbles; and realize automatic classification of the bubbles through the convolutional neural network, thereby improving the accuracy and efficiency of detection, helping to timely discover potential defects and take corresponding measures, thereby ensuring the quality of the brake pad and the driving safety of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 The present invention is a structural schematic diagram of a bubble imaging detection system for the inside of a brake pad. DETAILED DESCRIPTION
[0029] 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. 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.
[0030] Example 1
[0031] A bubble imaging detection system for the interior of a brake pad comprises a brake pad interior imaging module, a bubble recognition module, a bubble classification module and a user interface.
[0032] The brake pad internal imaging module in the embodiment of the present invention is connected to the bubble recognition module, the bubble recognition module is connected to the bubble classification module, and the bubble classification module is connected to the user interface.
[0033] The brake pad internal imaging module is used to obtain an image of the internal structure of the brake pad, including: a light source for emitting a light beam into the interior of the brake pad; an optical sensor for receiving the light beam reflected and transmitted through the interior of the brake pad and converting it into an electrical signal; and an image processing unit for converting the electrical signal of the optical sensor into image data to obtain bubble area parameters.
[0034] The brake pad internal imaging module in the embodiment of the present invention uses multi-spectral imaging technology to obtain different information of an object through light of different wavelengths to increase the penetration depth of light and help obtain more information inside the brake pad.
[0035] The bubble recognition module includes a bubble image enhancement unit, a bubble image segmentation unit and a bubble parameter calculation unit. The bubble image enhancement unit is connected to the bubble image segmentation unit, and the bubble image segmentation unit is connected to the bubble parameter calculation unit, and is used to identify and analyze the image data obtained by the imaging module inside the brake pad.
[0036] The bubble image enhancement unit in the embodiment of the present invention is used to enhance the image acquired by the imaging module inside the brake pad and transmit it to the bubble image segmentation unit. The enhancement formula is:
[0037]
[0038] Where E(x,y) is the enhanced image, I(x,y) is the original image, G(x,y) is the gradient value of image I(x,y) at position (x,y), and k is the adjustment coefficient.
[0039] The bubble image segmentation unit in the embodiment of the present invention is used to segment the enhanced image. When the pixel value of the enhanced image at the position (x, y) is greater than the threshold, the position is identified as a bubble area and marked as 1; when the pixel value of the enhanced image at the position (x, y) is less than or equal to the threshold, the position is identified as a non-bubble area and marked as 0; a binary image is obtained, wherein the bubble recognition formula is:
[0040]
[0041] Where B(x,y) is a binary image, where 1 represents the bubble area and 0 represents the non-bubble area. E(x,y) is the pixel value of the enhanced image at position (x, y), and T is the set threshold.
[0042] The bubble parameter calculation unit in the embodiment of the present invention identifies the bubble area in the binary image through image connected domain analysis technology, obtains the bubble area parameters of each bubble, including the area and perimeter of each bubble, and constructs a bubble relative number index evaluation model, a bubble morphology characteristic index evaluation model and a bubble distribution density evaluation model.
[0043] The formula of the bubble relative number index evaluation model in the embodiment of the present invention is:
[0044]
[0045] Where N b is the bubble relative number index evaluation value, M is the number of pixels of the image in the vertical direction, N is the number of pixels of the image in the horizontal direction, B(x,y) is a binary image, where B(x,y)=1 represents the bubble area, and B(x,y)=0 represents the non-bubble area.
[0046] The formula of the bubble morphology characteristic index evaluation model in the embodiment of the present invention is:
[0047]
[0048] Where S is the evaluation value of the bubble morphological characteristic index, M is the number of pixels of the image in the vertical direction, N is the number of pixels of the image in the horizontal direction, B(x,y) is a binary image, where B(x,y)=1 represents the bubble area, B(x,y)=0 represents the non-bubble area, A(x,y) is the area of the bubble at position (x,y), and P(x,y) is the circumference of the bubble at position (x,y).
[0049] The formula of the bubble density evaluation model in the embodiment of the present invention is:
[0050]
[0051] Where D is the estimated value of the bubble distribution density, M is the number of pixels in the vertical direction of the image, N is the number of pixels in the horizontal direction of the image, and N is the number of pixels in the horizontal direction of the image. b is the total number of bubble regions in the image, A i is the area of the ith bubble, x i is the horizontal coordinate of the center position of the ith bubble, y i is the ordinate of the center position of the i-th bubble, A max is the maximum area among all bubbles.
[0052] The bubble recognition module is used to build a bubble classification model based on the convolutional neural network, and use the image data of the bubbles inside the brake pad and the annotation information of the bubble area to identify the probability distribution of the bubbles in the corresponding category labels and obtain the bubble classification results.
[0053] The bubble classification module is based on the enhanced image and the segmented bubble area, and inputs the constructed convolutional neural network to obtain the probability distribution of each bubble and the corresponding category label. The final bubble classification result is obtained by taking the category label corresponding to the maximum probability. The category labels include structural bubbles, stress concentration bubbles, uniformly distributed bubbles, tiny isolated bubbles and large-sized aggregated bubbles.
[0054] The structural bubbles in the embodiment of the present invention are bubbles located in the core area of the brake pad, have a high relative quantity index, regular shape, and high distribution density, and affect the overall structural strength and durability of the brake pad.
[0055] The stress concentration bubbles in the embodiment of the present invention represent bubbles located at the edge or key stress points, with irregular shapes and high distribution density, forming stress concentration points, which are prone to cause local damage or cracks in the brake pad.
[0056] The uniformly distributed bubbles in the embodiment of the present invention are uniformly distributed throughout the brake pad, with a medium relative quantity index and a medium distribution density, which reduce the overall strength of the brake pad but do not cause stress concentration bubbles.
[0057] The tiny isolated bubbles in the embodiment of the present invention are bubbles located in non-critical areas, with small area, small perimeter, low relative number index, sparse distribution density, and little impact on the overall performance.
[0058] The large-sized aggregated bubbles in the embodiment of the present invention are bubbles located in a local area, with a large area, a large perimeter, a high relative number index, and a high distribution density, which affect the local performance and may cause the brake pad to fail.
[0059] The steps of constructing a convolutional neural network in the embodiment of the present invention are:
[0060] Step S1, determining that the input layer includes image data of bubbles inside the brake pad and annotation information of the bubble area, wherein the image data includes an original image, an enhanced image, and a segmented bubble area image, and the annotation information includes a relative quantity index evaluation value, a morphological feature index evaluation value, a distribution density evaluation value, a position coordinate, an area, and a perimeter of the bubble;
[0061] Step S2, cleaning and standardizing the above data to extract image features and bubble feature vectors;
[0062] Step S3, dividing the historical data into a training set and a validation set, using the training set to train a convolutional neural network algorithm model, and building a bubble classification model;
[0063] Step S4, after the training is completed, the bubble classification model is verified using the verification set;
[0064] Step S5, in the new bubble detection task, the enhanced image and the segmented bubble area image are input, the bubble classification model is used for prediction, and the corresponding bubble classification result is output.
[0065] The following is a Python code example that uses a convolutional neural network to build a bubble classification model to classify bubbles. Please note that this example is only a starting point and may need to be adjusted according to actual conditions and device interfaces in actual applications.
[0066]
[0067]
[0068]
[0069] This code is only used as an example. In actual applications, appropriate modifications and adjustments need to be made according to specific circumstances.
[0070] The user interface is used to display the image data of the bubbles inside the brake pad and the annotation information of the bubble area output by the brake pad internal imaging module and the bubble recognition module, as well as the bubble classification results output by the bubble classification module.
[0071] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
[0072] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A bubble imaging detection system for the inside of a brake pad, comprising a brake pad internal imaging module, a bubble recognition module, a bubble classification module and a user interface, It is characterized in that The internal imaging module of the brake pad is connected to the bubble recognition module, the bubble recognition module is connected to the bubble classification module, and the bubble classification module is connected to the user interface; the bubble recognition module includes a bubble image enhancement unit, a bubble image segmentation unit and a bubble parameter calculation unit, the bubble image enhancement unit is connected to the bubble image segmentation unit, and the bubble image segmentation unit is connected to the bubble parameter calculation unit, which is used to identify and analyze the image data obtained by the internal imaging module of the brake pad, obtain the bubble area parameters, and build a bubble distribution density evaluation model based on the acquired parameters; the bubble recognition module is used to build a bubble classification model based on a convolutional neural network, and use the image data of the bubbles inside the metal-free friction material brake pad and the annotation information of the bubble area to identify the probability distribution of the bubbles in the corresponding category label and obtain the bubble classification result; wherein, the formula of the bubble distribution density evaluation model is: Where D is the estimated value of the bubble distribution density, M is the number of pixels in the vertical direction of the image, N is the number of pixels in the horizontal direction of the image, and N is the number of pixels in the horizontal direction of the image. b is the total number of bubble regions in the image, A i is the area of the ith bubble, x i is the horizontal coordinate of the center position of the ith bubble, y i is the ordinate of the center position of the i-th bubble, A max is the maximum area among all bubbles.
2. A bubble imaging detection system for the inside of a brake pad according to claim 1, It is characterized in that The bubble image enhancement unit is used to enhance the image obtained by the imaging module inside the brake pad and transmit it to the bubble image segmentation unit; the bubble image segmentation unit is used to segment the enhanced image. When the pixel value of the enhanced image at the position (x, y) is greater than the threshold, the position is determined to be a bubble area and marked as 1; when the pixel value of the enhanced image at the position (x, y) is less than or equal to the threshold, the position is determined to be a non-bubble area and marked as 0; a binary image is obtained.
3. The bubble imaging detection system for the inside of a brake pad according to claim 1, It is characterized in that The bubble parameter calculation unit identifies the bubble area in the binary image through the image connected domain analysis technology, obtains the bubble area parameters of each bubble, including the area and perimeter of each bubble, and constructs a bubble relative number index evaluation model and a bubble morphological characteristic index evaluation model, wherein the formula of the bubble relative number index evaluation model is: Where N b is the bubble relative number index evaluation value, M is the number of pixels of the image in the vertical direction, N is the number of pixels of the image in the horizontal direction, B(x,y) is a binary image, where B(x,y)=1 represents the bubble area, and B(x,y)=0 represents the non-bubble area; The formula of the bubble morphology characteristic index evaluation model is: Where S is the evaluation value of the bubble morphological characteristic index, M is the number of pixels of the image in the vertical direction, N is the number of pixels of the image in the horizontal direction, B(x,y) is a binary image, where B(x,y)=1 represents the bubble area, B(x,y)=0 represents the non-bubble area, A(x,y) is the area of the bubble at position (x,y), and P(x,y) is the circumference of the bubble at position (x,y).
4. The air bubble imaging detection system for the inside of a brake pad according to claim 1, It is characterized in that The bubble classification module is based on the enhanced image and the segmented bubble area, and inputs the constructed convolutional neural network to obtain the probability distribution of each bubble and the corresponding category label. The final bubble classification result is obtained by taking the category label corresponding to the maximum probability. The category labels include Class I bubbles, Class II bubbles, Class III bubbles, Class IV bubbles and Class V bubbles.
5. A bubble imaging detection system for the inside of a brake pad according to claim 1 or 4, It is characterized in that The steps to build a convolutional neural network are: Step S1, determining that the input layer includes image data of bubbles inside a low-metal friction material brake pad and annotation information of the bubble area, wherein the image data includes an original image, an enhanced image, and a segmented bubble area image, and the annotation information includes a relative quantity index evaluation value, a morphological feature index evaluation value, a distribution density evaluation value, a position coordinate, an area, and a perimeter of the bubble; Step S2, cleaning and standardizing the above data to extract image features and bubble feature vectors; Step S3, dividing the historical data into a training set and a validation set, using the training set to train a convolutional neural network algorithm model, and building a bubble classification model; Step S4, after the training is completed, the bubble classification model is verified using the verification set; Step S5, in the new bubble detection task, the enhanced image and the segmented bubble area image are input, the bubble classification model is used for prediction, and the corresponding bubble classification result is output.
6. The bubble imaging detection system for the inside of a brake pad according to claim 1, It is characterized in that The brake pad internal imaging module is used to obtain an image of the internal structure of the low-metal friction material brake pad, and includes: a light source for emitting a light beam into the interior of the low-metal friction material brake pad; an optical sensor for receiving the light beam reflected and transmitted through the interior of the low-metal friction material brake pad and converting it into an electrical signal; and an image processing unit for converting the electrical signal of the optical sensor into image data.
7. The air bubble imaging detection system for the inside of a brake pad according to claim 1, It is characterized in that The user interface is used to display the image data of bubbles inside the low-metal friction material brake pad output by the brake pad internal imaging module and the bubble recognition module and the annotation information of the bubble area, as well as the bubble classification results output by the bubble classification module.
