A nondestructive testing method for internal defects of rubber sheets based on light transmission detection

Abstract

The present invention discloses a nondestructive testing method for internal defects of rubber plates based on light transmission detection, which relates to the technical field of quality inspection, and includes the following steps: S1, conducting adaptability test; S2, adjusting the matching degree of light source and video acquisition device; S3, determining the detection path; S4, performing layered processing to distinguish surface and internal defect features; S5, performing regional calibration and classification; S6, performing spectral verification on the classified defective areas; S7, performing physical verification on the calibrated defective areas. The nondestructive testing method for internal defects of rubber plates can effectively detect hidden defects such as pores, impurities, joints and adhesions inside rubber plates and their products by optimizing the matching degree of light source and video acquisition device, multi-angle dynamic path planning, adaptive attention mechanism, and spectral verification and classification calibration of defects, generate accurate defect distribution map and realize dynamic optimization, ensure the consistency and reliability of product quality, and meet high-standard industrial needs.

Description

Technical Field

[0001] The invention relates to the technical field of quality inspection, and in particular to a nondestructive inspection method for internal defects of a rubber plate based on light transmission inspection. Background Art

[0002] During the production process, rubber sheets and related products (such as cloth-reinforced rubber sheets and tape products) are prone to internal hidden defects such as pores, impurities, joints, and adhesion due to the quality of raw materials, the complexity of molding and vulcanization processes. These defects are often hidden between layers or in cross-sections of rubber materials and are difficult to detect through conventional surface inspections. However, in actual use, these defects may lead to stress concentration, which in turn causes material damage or failure, posing a major hidden danger to product performance and safety. In order to achieve high-precision non-destructive testing, light transmission detection technology combined with video acquisition and intelligent analysis methods is the focus of research. Using the strong light transmission characteristics to reveal the internal structure of rubber and identifying anomalies through high-resolution imaging and in-depth analysis has become an important direction of modern industrial inspection.

[0003] Although the existing light transmission detection method can detect internal defects of rubber sheets to a certain extent, it has the following shortcomings:

[0004] Insufficient adaptability to complex structures: Existing path scanning methods are mostly fixed linear scanning, which is difficult to fully cover rubber sheets with laminated structures or complex geometric shapes, and is prone to blind spots.

[0005] Lack of intelligent classification and calibration: Many methods rely on simple image threshold segmentation and cannot accurately distinguish different types of defects, resulting in high false detection and missed detection rates. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a nondestructive testing method for internal defects of a rubber plate based on light transmission detection to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides a nondestructive testing method for internal defects of a rubber plate based on light transmission detection, comprising the following steps:

[0009] S1. Conduct adaptability test before light transmission test of rubber sheet;

[0010] S2. Based on the adaptability test results, adjust the matching degree between the light source and the video acquisition device;

[0011] S3. After the adjustment is completed, the detection path is determined to cover the entire surface and deep areas of the target sample;

[0012] S4, based on the data obtained by scanning the detection path, performing layered processing to distinguish surface and internal defect features;

[0013] S5. After the defect extraction is completed, the area is demarcated and classified;

[0014] S6. Perform spectral verification on the classified defect areas to distinguish the defect types and their causes;

[0015] S7. Based on the results of step S5 and step S6, physically verify the calibrated defect area to confirm the detection accuracy.

[0016] To further optimize the technical solution, in step S1, the adaptability test includes:

[0017] The light transmittance of the rubber sheet is measured by an optical transmittance testing device using a standardized light source such as an LED array light source or a laser light source;

[0018] According to the thickness and color differences of the rubber sheet, adjust the wavelength range and luminous intensity of the light source to ensure that the light can penetrate the middle layer of the rubber sheet and provide a clear defect image;

[0019] The video acquisition device should also be debugged to avoid missed or false detection of defects due to improper parameter selection.

[0020] To further optimize the technical solution, in step S2, adjusting the matching degree between the light source and the video acquisition device includes:

[0021] Combined with the light transmittance of the rubber sheet, the spatial angle and distance between the light source and the video device are adjusted;

[0022] The light source is low-angle oblique light to highlight the light and shadow contrast of internal defects and assist the background light source to eliminate surface reflection interference;

[0023] The video acquisition device uses a high-resolution camera with a CMOS or CCD sensor and is equipped with a lens with a continuous zoom function;

[0024] Through real-time image feedback, the focus accuracy and resolution parameters of the video acquisition device are optimized to ensure that the edge features and light and shadow changes of tiny defects are captured during inspection.

[0025] To further optimize the technical solution, in step S3, the determination of the detection path includes:

[0026] Combined with the shape and size of the rubber sheet, write an automated inspection path program and use a multi-axis control platform with XY axis or a six-degree-of-freedom robotic arm to perform dynamic sample scanning;

[0027] During path planning, the distance between the light source and the sample is kept constant, and a special path is designed for the laminated parts of the rubber sheet to enhance the light source transmission effect and avoid missing defects in key areas.

[0028] To further optimize the technical solution, the special path of the stacking part is designed to include:

[0029] Light source transmission enhancement path at the stacked part;

[0030] Key area focused scan path;

[0031] Dynamic paths that adapt to material deformation;

[0032] Local detail enhancement path;

[0033] Time-optimized path partitioning scanning: prioritize dense scanning paths for critical areas at the center and edges of the stack, and fast scanning paths for non-critical areas.

[0034] To further optimize the technical solution, the light source transmission enhancement path of the stacked part includes:

[0035] Multi-angle scanning path: multiple tilt angles are set at the stacking part, including 30°, 45°, and 60° light source incident paths, which are used to enhance the scattering and transmission effects of light in the material by changing the direction of the incident light;

[0036] The multi-axis control platform controls the camera and light source to move simultaneously, continuously capturing images from different angles;

[0037] Multiple local scans of the overlapping area of ​​the stack: With the stacked area as the center, a local path of multiple scans is designed. The luminous intensity of the light source increases step by step with each scan, and the focal length is adjusted to optimize the clarity of the stacked boundary. Multiple images obtained by overlapping scans are used for superposition analysis to improve the resolution of deep defects.

[0038] To further optimize the technical solution, the dynamic path adapted to material deformation includes:

[0039] Equipped with laser distance measuring sensors or optical profile scanning equipment, it can sense the slight deformation of the rubber sheet surface in real time, dynamically adjust the Z-axis height of the scanning path, and maintain the optimal distance between the light source and the rubber sheet;

[0040] In the stacking parts where the bending angle is greater than 5°, a nonlinear path based on the material curvature is planned, and the light source and camera move synchronously along the curved surface to ensure uniform light penetration.

[0041] Further optimizing the technical solution, in the step S5, a defect classification model is constructed to calibrate and classify the internal defects of the rubber sheet, and the defect classification model includes an input layer, a convolutional layer, an attention module, a feature classifier and a dynamic calibration module;

[0042] The input layer and convolution layer are used for input and convolution feature extraction;

[0043] The attention module is used to design an adaptive attention mechanism so that the model focuses on the feature representation of the salient area;

[0044] The feature classifier predicts defects using a fully connected layer and a classification function;

[0045] The dynamic calibration module combines the classification results and the spatial location information to generate a defect distribution map.

[0046] Further optimize the technical solution, in the input layer and the convolution layer, the input image ,high ,width , Number of channels It is sent to the initial convolution layer to extract low-level features; the convolution process is as follows:

[0047] ;

[0048] in,

[0049] : Low-level feature map;

[0050] and : Convolution kernel weights and biases;

[0051] : convolution operator;

[0052] : activation function;

[0053] Stack multiple deep convolutional layers to gradually extract high-level shape and texture features;

[0054] In the attention module, the adaptive attention weight calculation is as follows:

[0055] ;

[0056] in,

[0057] : By feature and reference features Calculate the attention score, is the weight matrix, Represents feature splicing;

[0058] : No. In the feature map The attention weight of each position;

[0059] The generated attention features are:

[0060] ;

[0061] In the feature classifier, the defect prediction is as follows:

[0062] ;

[0063] in,

[0064] :category The predicted probability distribution of

[0065] and : classifier weights and biases;

[0066] : Feature map processed by the attention mechanism;

[0067] Defect categories include pores, impurities, joints and bonding, with the maximum pass probability:

[0068] ;

[0069] The dynamic calibration module generates a defect distribution diagram as follows:

[0070] ;

[0071] in, : Position in defect distribution diagram The calibration value of

[0072] Dynamic calibration aligns the classification results with the physical locations based on the mapping relationship between the feature map and the input image, and outputs a defect location map with category labels.

[0073] To further optimize the technical solution, in step S6, a multi-band scanning is performed on the calibrated defect area using a hyperspectral imaging technique to extract the optical absorption and reflection characteristics of the material;

[0074] By analyzing the spectral curves in different bands, the physical properties and chemical composition of the defects can be determined.

[0075] In a second aspect, an embodiment of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of a method for non-destructive testing of internal defects of a rubber plate based on light transmission detection as described in the first aspect of the present invention are implemented.

[0076] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of a method for non-destructive testing of internal defects of a rubber plate based on light transmission detection as described in the first aspect of the present invention are implemented.

[0077] Compared with the prior art, the present invention provides a nondestructive testing method for internal defects of rubber plates based on light transmission detection, which has the following beneficial effects:

[0078] This nondestructive testing method for internal defects of rubber sheets based on light transmission detection solves the shortcomings of existing technologies in terms of adaptability to complex structures, detection accuracy, and intelligent classification by optimizing the matching of light sources and video acquisition devices, multi-angle dynamic path planning, adaptive attention mechanism, and spectral verification and classification calibration of defects. This method can effectively detect hidden defects such as pores, impurities, joints, and adhesion inside rubber sheets and their products, generate accurate defect distribution maps, and achieve dynamic optimization, ensuring the consistency and reliability of product quality and meeting high-standard industrial needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0080] Figure 1 A schematic diagram of a flow chart of a nondestructive testing method for internal defects of a rubber plate based on light transmission detection proposed by the present invention;

[0081] Figure 2 A schematic diagram of the design process of a special path of a laminated part in a nondestructive testing method for internal defects of a rubber plate based on light transmission detection proposed by the present invention;

[0082] Figure 3 The present invention provides a flow chart of a defect classification model in a nondestructive testing method for internal defects of a rubber plate based on light transmission detection. DETAILED DESCRIPTION

[0083] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0084] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0085] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0086] Embodiment 1:

[0087] Reference Figure 1 to Figure 3 , which is the first embodiment of the present invention, provides a nondestructive testing method for internal defects of a rubber sheet based on light transmission detection, comprising the following steps:

[0088] S1. Adaptability test before light transmission test of rubber sheet

[0089] In this embodiment, for the material of the rubber sheet, such as natural rubber, silicone rubber and other translucent synthetic rubber, the adaptability test includes:

[0090] The light transmittance of the rubber sheet is measured by an optical transmittance testing device using a standardized light source such as an LED array light source or a laser light source;

[0091] According to the thickness and color differences of the rubber sheet, adjust the wavelength range and luminous intensity of the light source to ensure that the light can penetrate the middle layer of the rubber sheet and provide a clear defect image;

[0092] The video acquisition device should also be debugged to avoid missed or false detection of defects due to improper parameter selection.

[0093] S2. Based on the adaptability test results, adjust the matching degree of the light source and the video acquisition device

[0094] In this embodiment, adjusting the matching degree between the light source and the video acquisition device includes:

[0095] Combined with the light transmittance of the rubber sheet, the spatial angle and distance between the light source and the video device are adjusted;

[0096] The light source is low-angle oblique light to highlight the light and shadow contrast of internal defects and assist the background light source to eliminate surface reflection interference;

[0097] The video acquisition device uses a high-resolution camera with a CMOS or CCD sensor and is equipped with a lens with a continuous zoom function;

[0098] Through real-time image feedback, the focus accuracy and resolution parameters of the video acquisition device are optimized to ensure that the edge features and light and shadow changes of tiny defects are captured during inspection.

[0099] S3. After adjustment, determine the detection path to cover the entire surface and deep areas of the target sample

[0100] In this embodiment, the determination of the detection path includes:

[0101] Combined with the shape and size of the rubber sheet, write an automated inspection path program and use a multi-axis control platform with XY axis or a six-degree-of-freedom robotic arm to perform dynamic sample scanning;

[0102] Path planning needs to take into account the overlapping areas of the samples to avoid the existence of scanning blind areas, while ensuring that the distance between the light source and the sample is constant. Special paths are designed for the laminated parts of the rubber sheets to enhance the light source transmission effect and avoid missing defects in key areas.

[0103] Furthermore, the special path of the stacking part includes:

[0104] Light source transmission enhancement path at the stacked part

[0105] For laminated rubber sheets, the light transmittance in the laminated area is usually low, which makes internal defect detection more difficult. Therefore, path planning requires enhanced detection in these areas.

[0106] Multi-angle scanning path: multiple tilt angles are set at the stacking part, including 30°, 45°, and 60° light source incident paths, which are used to enhance the scattering and transmission effects of light in the material by changing the direction of the incident light;

[0107] The multi-axis control platform controls the camera and light source to move simultaneously, continuously capturing images from different angles;

[0108] Multiple local scans of the overlapping area of ​​the stack: With the stacked area as the center, a local path of multiple scans is designed. The luminous intensity of the light source increases step by step with each scan, and the focal length is adjusted to optimize the clarity of the stacked boundary. Multiple images obtained by overlapping scans are used for superposition analysis to improve the resolution of deep defects.

[0109] Key area focused scan path

[0110] Since there may be poor bonding or pores at the laminate joints, the detection accuracy of this area needs to be enhanced.

[0111] Prioritize scanning sequences for critical areas in the inspection path (e.g., laminate edges, bond seams), where the system can optionally slow down scanning speeds and increase resolution.

[0112] The key areas are scanned in a spiral path, that is, gradually expanding from the center of the joint outward to ensure that the entire joint area is covered with high-resolution images.

[0113] Dynamic paths that adapt to material deformation

[0114] The rubber sheet in the laminated area may be slightly deformed or bent during the manufacturing process, and the standard path may not completely conform to its surface.

[0115] Equipped with laser distance measuring sensors or optical profile scanning equipment, it can sense the slight deformation of the rubber sheet surface in real time, dynamically adjust the Z-axis height of the scanning path, and maintain the optimal distance between the light source and the rubber sheet;

[0116] In the stacking parts where the bending angle is greater than 5°, a nonlinear path based on the material curvature is planned, and the light source and camera move synchronously along the curved surface to ensure uniform light penetration.

[0117] Local detail enhancement path

[0118] During regular scanning, the system may detect certain optical anomalies but cannot accurately identify the defect type or size, requiring local detail inspection.

[0119] On the basis of the conventional scanning path, a dynamic feedback mechanism is added. When a suspected defect area is detected, the system triggers the detail enhancement mode and performs high-density path coverage on the area.

[0120] Detailed scanning can adopt a grid path, with the light source scanning the entire suspected defect area point by point, and the resolution is increased to 2-3 times that of conventional scanning to obtain more image information.

[0121] Time-optimized path partition scanning

[0122] In order to balance the detection efficiency and accuracy, the rubber sheet can be divided into priority areas and regular areas, and different scanning strategies are adopted for different areas.

[0123] Based on historical inspection data, dense scanning paths are prioritized for areas with high defect incidence, such as stack centers and edges.

[0124] Use fast scanning paths for low-risk areas to reduce unnecessary inspection time, while avoiding lag or jitter from frequent device movements when switching paths.

[0125] The above special path design ensures that the internal defects of the laminated rubber sheet can be fully detected, while taking into account both efficiency and accuracy, reducing the risk of missed detection or misdetection.

[0126] S4. Based on the data obtained by scanning the detection path, layer processing is performed to distinguish surface and internal defect features

[0127] In this embodiment, the collected continuous images are processed frame by frame, and the optical information of different depth levels of the rubber sheet is extracted by using an image layering algorithm (such as optical tomography or deep learning segmentation model). The image of each layer is filtered to enhance the edge details of the internal structure, and at the same time, specific grayscale differences are used to detect defects such as pores, impurities and joints. By focusing on light and shadow distortion and reflection anomalies, the position and size of defects between layers are determined, providing a reference for subsequent verification links.

[0128] S5. After defect extraction, perform area calibration and classification

[0129] In this embodiment, a defect classification model is constructed to calibrate and classify internal defects of the rubber sheet, and the defect classification model includes an input layer, a convolutional layer, an attention module, a feature classifier, and a dynamic calibration module;

[0130] The input layer and convolution layer are used for input and convolution feature extraction;

[0131] The attention module is used to design an adaptive attention mechanism so that the model focuses on the feature representation of the salient area;

[0132] The feature classifier predicts defects using a fully connected layer and a classification function;

[0133] The dynamic calibration module combines the classification results and the spatial location information to generate a defect distribution map.

[0134] Furthermore, in the input layer and the convolution layer, the input image ,high ,width , Number of channels It is sent to the initial convolution layer to extract low-level features; the convolution process is as follows:

[0135] ;

[0136] in,

[0137] : Low-level feature map;

[0138] and : Convolution kernel weights and biases;

[0139] : convolution operator;

[0140] : activation function.

[0141] Multiple deep convolutional layers are stacked to gradually extract high-level shape and texture features:

[0142] ;

[0143] in, : No. The feature map of the layer, is the total number of layers.

[0144] These layers will gradually capture important features such as edge information of defects, light and shadow contrast, etc.

[0145] In the attention module, the adaptive attention weight calculation is as follows:

[0146] ;

[0147] in,

[0148] : By feature and reference features Calculate the attention score, is the weight matrix, Represents feature splicing;

[0149] : No. In the feature map The attention weight of each position;

[0150] The generated attention features are:

[0151] ;

[0152] The attention module enhances the model’s perception of critical defect areas.

[0153] In the feature classifier, based on the extracted high-level features, a fully connected layer and a classification function are used to predict defects:

[0154] ;

[0155] in,

[0156] :category The predicted probability distribution of

[0157] and : classifier weights and biases;

[0158] : Feature map processed by the attention mechanism;

[0159] Defect categories include pores, impurities, joints and bonding, with the maximum pass probability:

[0160] ;

[0161] The dynamic calibration module combines the classification results and spatial location information to generate a defect distribution map as shown below:

[0162] ;

[0163] in, : Position in defect distribution diagram The calibration value of

[0164] Dynamic calibration aligns the classification results with the physical locations based on the mapping relationship between the feature map and the input image, and outputs a defect location map with category labels.

[0165] In summary, when the model is used, it includes the following processes:

[0166] Input image preprocessing: Normalize the original images collected by the optical detection system to a standard size to ensure that the dimensions of the data input model are consistent (for example, unified to a size of 256×256).

[0167] Feature extraction and enhancement: The convolutional layers and adaptive attention mechanism are used to extract and enhance the significant features of the defect area, such as edge discontinuities and abnormal light and shadow contrast.

[0168] Classification prediction: In the classification head, the extracted features are predicted to be classified (such as pores, impurities, etc.), and the probability value of each category is output through the formula.

[0169] Defect labeling: The prediction results and location information are combined to generate a defect distribution map, where each defect area is annotated with a bounding box and additional category information (for example, pores are marked with a red box and impurities are marked with a green box).

[0170] Dynamic feedback optimization: For detected complex areas, secondary detection is performed by re-adjusting the acquisition parameters to reduce the possibility of false detection or missed detection.

[0171] S6. Perform spectral verification on the classified defect areas to distinguish the defect types and their causes

[0172] In this embodiment, a hyperspectral imaging technique is used to perform multi-band scanning on the calibrated defect area to extract the optical absorption and reflection characteristics of the material;

[0173] By analyzing the spectral curves in different bands, the physical properties (such as density differences) and chemical composition (such as the source of impurities) of the defects can be determined. For example, the spectral reflectivity of pores is lower than that of the surrounding areas, and the spectral absorption characteristics of impurities may indicate specific material composition. This step combines spectral analysis with raw image data to ensure the accuracy of classification results and provide feedback for process improvement.

[0174] S7: Combine the results of step S5 and step S6 to physically verify the calibrated defect area to confirm the detection accuracy.

[0175] In this embodiment, the defect calibration area is taken out from the test sample, and the section is dissected by precision cutting equipment to observe the internal structure of the rubber. The dissected area is carefully observed with an optical microscope or a scanning electron microscope to verify the consistency between the morphology of the test defect and the actual position. This link pays special attention to the matching degree between the spectral analysis results and the dissection results to further optimize the test parameters.

[0176] Comprehensively evaluate test data and profile analysis results to identify the main causes of defects (such as raw material problems or abnormal vulcanization process parameters).

[0177] Embodiment 2:

[0178] This embodiment also provides a computer device, which is applicable to a method for nondestructive testing of internal defects of a rubber plate based on light transmission detection, and includes a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to implement a method for nondestructive testing of internal defects of a rubber plate based on light transmission detection as proposed in the above embodiment.

[0179] This embodiment further provides a storage medium on which a computer program is stored. When the program is executed by a processor, a nondestructive testing method for internal defects of a rubber plate based on light transmission detection as proposed in the above embodiment is implemented.

[0180] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0181] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0182] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0183] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk case (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0184] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit with a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit with a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0185] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A nondestructive testing method for internal defects of rubber sheets based on light transmission detection, characterized in that: The following steps are involved: S1. Conduct adaptability test before light transmission test of rubber sheet; S2. Based on the adaptability test results, adjust the matching degree between the light source and the video acquisition device; S3. After the adjustment is completed, the detection path is determined to cover the entire surface and deep areas of the target sample; Determination of the detection path, including: Combined with the shape and size of the rubber sheet, write an automated inspection path program and use a multi-axis control platform with XY axis or a six-degree-of-freedom robotic arm to perform dynamic sample scanning; When planning the path, the distance between the light source and the sample is kept constant. A special path is designed for the laminated part of the rubber sheet to enhance the light source transmission effect and avoid missing defects in key areas. Special paths for stacking parts are designed including: Light source transmission enhancement path at the stacked part; Key area focused scan path; Dynamic paths that adapt to material deformation; Local detail enhancement path; Time-optimized path partition scanning: prioritize dense scanning paths for critical areas at the center and edge of the stack, and fast scanning paths for non-critical areas; S4, based on the data obtained by scanning the detection path, performing layered processing to distinguish surface and internal defect features; S5. After the defect extraction is completed, the area is demarcated and classified; S6. Perform spectral verification on the classified defect areas to distinguish the defect types and their causes; S7. Based on the results of step S5 and step S6, physically verify the calibrated defect area to confirm the detection accuracy.

2. The nondestructive testing method for internal defects of rubber sheets based on light transmission detection according to claim 1 is characterized in that: In step S1, the adaptability test includes: The light transmittance of the rubber sheet is measured by an optical transmittance testing device using a standardized light source such as an LED array light source or a laser light source; According to the thickness and color differences of the rubber sheet, adjust the wavelength range and luminous intensity of the light source to ensure that the light can penetrate the middle layer of the rubber sheet and provide a clear defect image; The video acquisition device should also be debugged to avoid missed or false detection of defects due to improper parameter selection.

3. The nondestructive testing method for internal defects of rubber sheets based on light transmission detection according to claim 1, characterized in that: In step S2, adjusting the matching degree between the light source and the video acquisition device includes: Combined with the light transmittance of the rubber sheet, the spatial angle and distance between the light source and the video device are adjusted; The light source is low-angle oblique light to highlight the light and shadow contrast of internal defects and assist the background light source to eliminate surface reflection interference; The video acquisition device uses a high-resolution camera with a CMOS or CCD sensor and is equipped with a lens with a continuous zoom function; Through real-time image feedback, the focus accuracy and resolution parameters of the video acquisition device are optimized to ensure that the edge features and light and shadow changes of tiny defects are captured during inspection.

4. The nondestructive testing method for internal defects of rubber sheets based on light transmission detection according to claim 1, characterized in that: The light source transmission enhancement path of the stacked part includes: Multi-angle scanning path: multiple tilt angles are set at the stacking part, including 30°, 45°, and 60° light source incident paths, which are used to enhance the scattering and transmission effects of light in the material by changing the direction of the incident light; The multi-axis control platform controls the camera and light source to move simultaneously, continuously capturing images from different angles; Multiple local scans of the overlapping area of ​​the stack: With the stacked area as the center, a local path of multiple scans is designed. The luminous intensity of the light source increases step by step with each scan, and the focal length is adjusted to optimize the clarity of the stacked boundary. Multiple images obtained by overlapping scans are used for superposition analysis to improve the resolution of deep defects.

5. The nondestructive testing method for internal defects of rubber sheets based on light transmission detection according to claim 1, characterized in that: The dynamic path for adapting to material deformation includes: Equipped with laser distance measuring sensors or optical profile scanning equipment, it can sense the slight deformation of the rubber sheet surface in real time, dynamically adjust the Z-axis height of the scanning path, and maintain the optimal distance between the light source and the rubber sheet; In the stacking parts where the bending angle is greater than 5°, a nonlinear path based on the material curvature is planned, and the light source and camera move synchronously along the curved surface to ensure uniform light penetration.

6. The nondestructive testing method for internal defects of rubber sheets based on light transmission detection according to claim 1, characterized in that: In the step S5, a defect classification model is constructed to calibrate and classify the internal defects of the rubber sheet, wherein the defect classification model includes an input layer, a convolutional layer, an attention module, a feature classifier and a dynamic calibration module; The input layer and convolution layer are used for input and convolution feature extraction; The attention module is used to design an adaptive attention mechanism so that the model focuses on the feature representation of the salient area; The feature classifier predicts defects using a fully connected layer and a classification function; The dynamic calibration module combines the classification results and the spatial location information to generate a defect distribution map.

7. A nondestructive testing method for internal defects of a rubber sheet based on light transmission detection according to claim 6, characterized in that: In the input layer and convolution layer, the input image ,high ,width , Number of channels It is sent to the initial convolution layer to extract low-level features; the convolution process is as follows: ； in, : Low-level feature map; and : Convolution kernel weights and biases; : convolution operator; : activation function; Stack multiple deep convolutional layers to gradually extract high-level shape and texture features; In the attention module, the adaptive attention weight calculation is as follows: ； in, : By feature and reference features Calculate the attention score, is the weight matrix, Represents feature splicing; : No. In the feature map The attention weight of each position; The generated attention features are: ； In the feature classifier, the defect prediction is as follows: ； in, :category The predicted probability distribution of and : classifier weights and biases; : Feature map processed by the attention mechanism; Defect categories include pores, impurities, joints and bonding, with the maximum pass probability: ； The dynamic calibration module generates a defect distribution diagram as follows: ； in, : Position in defect distribution diagram The calibration value of Dynamic calibration aligns the classification results with the physical locations based on the mapping relationship between the feature map and the input image, and outputs a defect location map with category labels.

8. The nondestructive testing method for internal defects of rubber sheets based on light transmission detection according to claim 1, characterized in that: In the step S6, a multi-band scan is performed on the calibrated defect area using a hyperspectral imaging technique to extract the optical absorption and reflection characteristics of the material; By analyzing the spectral curves in different bands, the physical properties and chemical composition of the defects can be determined.

Images