Method and device for detecting appearance processing defects of adhesive tape

Abstract

The invention discloses an adhesive tape appearance processing defect detection method and device, and relates to the technical field of visual inspection.The detection method comprises the steps that a plurality of photographing point positions are arranged on an illumination detection area, and a plurality of detection images of an adhesive tape on the photographing point positions are obtained; comparing the detection image with a standard image to obtain a feature contour of the detection image; inputting the feature contour image into a defect detection model, and extracting a shooting parameter and an image parameter of the feature contour image; determining a plurality of groups of defect contrast data according to the shooting parameters of the feature contour image, and analyzing the similarity between the feature contour image and the defect contrast data according to the image parameters; and determining defect contrast data with the highest matching degree with the feature contour image according to the similarity, and outputting the defect type of the adhesive tape. The adhesive tape is polished, the defects of the adhesive tape are visually processed, and the defect type of the adhesive tape is determined based on the detection model, so that the reliability of adhesive tape defect detection can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of visual inspection, and in particular to a method and device for detecting appearance processing defects of a rubber strip. Background Art

[0002] During the production and preparation process of scraper rubber strips, processing operation defects can easily lead to processing appearance defects such as cuts, bubbles, and wear on the rubber strips, affecting the reliability of the use of the scraper rubber strips. The current appearance defect detection generally relies on manual inspection by staff. Since the rubber strips are transparent products, the appearance defect detection effect of manual inspection is insufficient, and the existing visual inspection equipment can only identify the defect position of the rubber strips, and cannot accurately identify the defect type of the rubber strips, affecting the quality and efficiency of the appearance inspection of the rubber strips. Summary of the invention

[0003] The purpose of the present invention is to overcome the shortcomings of the prior art. The present invention provides a method and device for detecting appearance processing defects of rubber strips. By lighting the rubber strips on the conveyor belt and detecting the defect type of the detection image in conjunction with a visual detection model, it helps to improve the quality of rubber strip detection and the efficiency of appearance defect detection.

[0004] The present invention provides a method for detecting appearance processing defects of a rubber strip, the detection method comprising:

[0005] Setting a plurality of photographing points on the light detection area, and obtaining a plurality of detection images of the adhesive strip at the plurality of photographing points;

[0006] Compare the detection image with the standard image to obtain the characteristic contour of the detection image;

[0007] Input the feature contour image into the defect detection model, and extract the shooting parameters and image parameters of the feature contour image;

[0008] Determine a plurality of groups of defect comparison data according to shooting parameters of the characteristic contour image, and analyze similarity between the characteristic contour image and the defect comparison data according to the image parameters;

[0009] Determine the defect comparison data with the highest matching degree with the characteristic contour image according to the similarity, and output the defect type of the rubber strip.

[0010] Furthermore, the step of setting a plurality of photographing points on the light detection area and obtaining a plurality of detection images of the adhesive strip at the plurality of photographing points includes:

[0011] A plurality of rubber strips are transported to a light detection area based on a conveyor belt, and the rubber strips are illuminated based on a light mechanism;

[0012] A plurality of photographing points are set on the light detection area, and the rubber strip after the light treatment is photographed and detected by a camera component to obtain a plurality of detection images of the rubber strip at the plurality of photographing points.

[0013] Furthermore, comparing the detection image with the standard image to obtain a characteristic contour of the detection image includes:

[0014] Preprocess the detection image and convert it into a pixel image;

[0015] The standard image of the rubber strip is extracted from the database, and after the standard image is converted into a pixel map, the pixel map of the standard image is subtracted from the pixel map of the detection image to extract the characteristic contour in the detection image.

[0016] Furthermore, the step of inputting the characteristic contour image into the defect detection model and extracting the shooting parameters and image parameters of the characteristic contour image includes:

[0017] Acquire shooting parameters of the characteristic contour image according to the detection image corresponding to the characteristic contour image, wherein the shooting parameters include a shooting angle and a shooting position;

[0018] Image parameters are extracted from the characteristic contour image, wherein the image parameters include shape and size data of the characteristic contour.

[0019] Furthermore, the determining of several groups of defect comparison data according to the shooting parameters of the characteristic contour image includes:

[0020] Extracting defect image data corresponding to the shooting parameters from a database of a detection model according to the shooting parameters of the characteristic contour image;

[0021] The defect image data is divided into several groups of defect comparison data according to the type of processing defects of the rubber strip appearance.

[0022] Furthermore, analyzing the similarity between the feature contour image and the defect comparison data according to the image parameters includes:

[0023] Screening a number of groups of defect comparison data according to the shape data of the characteristic contour image to obtain a number of groups of similar defect comparison data;

[0024] According to the size data of the characteristic contour image, a structural similarity analysis is performed on several groups of similar defect comparison data, and the analysis results are output.

[0025] Furthermore, the calculation formula for the structural similarity analysis is:

[0026]

[0027] Among them, SSIM(X,Y) is the similarity function, μx is the average value of X, μ y is the average value of Y, δ x 2 is the variance of X, δ y 2 is the variance of Y, δ xy is the covariance of X and Y, and C1 and C2 are constants.

[0028] Furthermore, the detection method also includes:

[0029] After confirming the defect type of the feature contour image, data enhancement is performed on the feature contour to obtain a data set of the feature contour;

[0030] The data set is divided into a training set and a test set, the defect detection model is trained based on the training set, and the model training result is verified by the test set.

[0031] The present invention also provides a device for detecting defects in the appearance of a rubber strip, wherein the device is used to execute the method for detecting defects in the appearance of a rubber strip;

[0032] The detection device comprises:

[0033] The transmission mechanism comprises a first transmission belt and a second transmission belt arranged in parallel, and a transparent adhesive film arranged between the first transmission belt and the second transmission belt, and a light detection area is arranged in the middle of the transmission mechanism;

[0034] Illumination mechanism: arranged in the illumination detection area, used for performing illumination treatment on the rubber strip;

[0035] Shooting component: used to shoot the rubber strip after lighting treatment and obtain the detection image of the rubber strip;

[0036] Main control system: used to process the inspection image of the rubber strip and obtain the appearance processing defects of the rubber strip.

[0037] Furthermore, a plurality of wave transition sections arranged obliquely at undulating positions are provided in the middle of the transmission mechanism, and the wave transition sections form the light detection area.

[0038] The present invention provides a method and device for detecting appearance processing defects of rubber strips. By lighting the rubber strips on a conveyor belt and cooperating with a detection model to perform visual defect detection on the detection image of the rubber strips, the defect image can be accurately analyzed and compared according to shooting parameters and image parameters, thereby improving the accuracy of defect type detection, and helping to improve the quality of rubber strip detection and the efficiency of appearance defect detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1is a flow chart of a method for detecting defects in the appearance of a rubber strip according to an embodiment of the present invention;

[0040] Figure 2 2 is a schematic diagram of the connection of a device for detecting defects in the appearance of a rubber strip according to an embodiment of the present invention;

[0041] Figure 3 Schematic diagram of the structure of a device for detecting defects in the appearance of rubber strips in an embodiment of the present invention. DETAILED DESCRIPTION

[0042] 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.

[0043] Embodiment 1:

[0044] Figure 1 A flow chart of a method for detecting appearance processing defects of a rubber strip according to an embodiment of the present invention is shown. The detection method comprises:

[0045] S11: setting a plurality of photographing points on the light detection area, and obtaining a plurality of detection images of the adhesive strip at the plurality of photographing points;

[0046] The method first uses a conveyor belt to transport the rubber strip to the illumination inspection area, and then uses an illumination mechanism to illuminate it. The illumination mechanism can use a ring light source, a backlight source, or a combined light source. The illumination intensity and angle can be dynamically adjusted according to the characteristics of the rubber strip to fully display the defect characteristics.

[0047] Next, the rubber strips after lighting treatment are photographed through multi-angle and multi-spectral camera components to obtain multiple inspection images. The camera component layout and image acquisition strategy have been optimized to capture different types of defect features. Optimized, based on the size characteristics of the rubber strips, in order to obtain higher shooting accuracy, a segmented shooting method can be adopted, that is, a detection image is composed of multiple smaller images. For example, the rubber strip is divided into 4 areas along the length direction, and images of the 4 areas are collected separately, and then the 4 images are spliced ​​into one inspection image, which can achieve more accurate image acquisition accuracy.

[0048] In practical applications, this method can adjust parameters according to the production line speed and the characteristics of the rubber strip. For example, the light intensity can be adjusted between 100-1000 lumens, the camera resolution can be selected as 4K or 8K, and the image acquisition frequency can be set to 30-60 frames per second. The 3D reconstruction accuracy can reach 0.1-0.5 mm.

[0049] The method of the embodiment of the present invention has significant advantages over traditional two-dimensional image analysis technology. By optimizing the illumination processing, the surface defect characteristics of the rubber strip can be better displayed and the detection sensitivity can be improved. In actual tests, the method can detect tiny defects as small as 0.05 mm in diameter. The combination of multi-angle image acquisition and three-dimensional reconstruction technology not only improves the accuracy of defect positioning, but also obtains depth information, which is crucial for evaluating the impact of defects.

[0050] In addition, the method of the embodiment of the present invention can complete high-precision defect detection without affecting the production line speed through real-time processing and analysis. In an actual case, the method was applied to a high-speed production line that processes 100 meters of rubber strips per minute, and the detection accuracy rate reached 99.5%, which is much higher than the 90% of traditional manual detection.

[0051] The method of the embodiment of the present invention performs well in processing transparent or translucent adhesive strips, and the optimized light processing technology effectively enhances the visibility of defects. At the same time, through efficient parallel processing and real-time analysis, high-precision detection can be achieved while maintaining high production efficiency. Finally, the provided three-dimensional visualization model can more intuitively display defect information, help make quick decisions, and improve production management efficiency.

[0052] Specifically, the rubber strip after the lighting treatment is photographed and inspected by the camera component to obtain a number of inspection images, including: photographing the rubber strip after the lighting treatment based on the camera component to obtain an initial image; inputting the initial image into the inspection system, and pre-processing the initial image through the inspection system to obtain a number of inspection images.

[0053] This technical solution obtains the test image through two steps: first, use the camera component to shoot the rubber strip after lighting treatment to obtain the initial image. This step ensures the integrity and accuracy of the surface information of the rubber strip. Secondly, the initial image is input into the detection system for preprocessing to obtain several test images. This step can optimize the initial image, such as denoising, contrast enhancement, etc., so that subsequent defect detection is more accurate. Through these two steps, the solution can obtain high-quality test images, providing a reliable data basis for subsequent defect feature extraction, defect type determination and three-dimensional reconstruction.

[0054] In the embodiment of the present invention, the camera assembly can use various types of cameras, such as high-speed cameras, industrial cameras or high-resolution cameras. These cameras can be selected according to the moving speed of the rubber strip and the detection requirements. For example, when the rubber strip moves at a fast speed, a high-speed camera can be selected to ensure that a clear image is captured; when it is necessary to detect tiny defects, a high-resolution camera can be selected.

[0055] After acquiring the initial image, preprocessing can include a variety of operations, such as image denoising, contrast enhancement, edge detection, etc. The selection and parameter settings of these operations can be adjusted according to the specific rubber strip material and the possible defect types.

[0056] S12: Compare the detection image with the standard image to obtain a feature contour of the detection image.

[0057] Preprocess the detection image and convert it into a pixel map. According to needs, various preprocessing operations can be performed on the image, such as grayscale conversion, resizing, filtering, etc. By reading the preprocessed detection image, the detection image is represented as a pixel matrix, and the pixel image of the detection image can be displayed through the cv2.imshow() function of OpenCV.

[0058] A standard image of the rubber strip is extracted from the database, and after the standard image is converted into a pixel image, a difference calculation is performed between the pixel image of the standard image and the pixel image of the detection image to extract the characteristic contour in the detection image, that is, after the standard image is converted into a pixel image, the pixel image of the detection image is superimposed on the pixel image of the standard image, and the pixels of the overlapping part of the two are deleted, thereby retaining the characteristic contour image of the detection image, that is, the defect position contour of the detection image.

[0059] Furthermore, a standard piece of rubber strip is placed in a conveyor belt, and an image of the rubber strip of a standard workpiece is taken as a standard image, so that the detection image can extract the defect contour according to the standard image.

[0060] S13: Input the feature contour image into the defect detection model, and extract the shooting parameters and image parameters of the feature contour image.

[0061] According to the detection image corresponding to the characteristic contour image, the shooting parameters of the characteristic contour image are obtained, and the shooting parameters include the shooting angle and the shooting position. The concave and convex positions are set on the conveyor belt so that the rubber strip can change its posture state when passing through the concave and convex positions during the transportation process. The detection images of the rubber strip in different posture states can be obtained by cooperating with the shooting component.

[0062] According to the type of processing defects of the rubber strip, the rubber strip in different posture states can correspondingly improve the appearance contour features of different types of processing defects, thereby improving the accuracy of image detection.

[0063] Furthermore, when the rubber strip is facing the camera assembly, it is equivalent to observing the appearance of the rubber strip in the forward direction. The forward shooting method can improve the accuracy of detecting appearance defects such as bubbles, impurities, oil rings, glue, flow marks, cracks, uneven sandblasting, etc. on the rubber strip.

[0064] Among them, bubbles: present a circular defect characteristic outline; impurities: present an irregular closed figure surrounding the impurities;

[0065] Oil droplets: present irregular closed figures (such as the initial state of oil drops on the water surface); Glue: there is an irregular layer of glue on the surface of the working surface;

[0066] Flow marks: the working surface is partially or entirely covered with straight lines extending in the same direction; Wire drawing: the surface presents relatively regular wavy lines;

[0067] When the appearance of the rubber strip is qualified, the material is sprayed on the surface of the rubber strip to form a frosted surface. Based on the frosted surface, crack defects after spraying and uneven sandblasting can be obtained.

[0068] When the rubber strip and the shooting direction of the camera assembly form an acute angle or an obtuse angle, the accuracy of detecting appearance defects such as crushing, creases, and deformation of the rubber strip can be improved.

[0069] Furthermore, the appearance characteristics of the defects are: bruises: the surface shows very tiny bumps (such as the effect of hammering a metal sheet multiple times); creases and deformations: such as the effect of scratching a piece of white paper.

[0070] Based on the rubber strips in different postures, the camera component can improve the defect detection of different observation surfaces and different positions of the rubber strips, thereby improving the comprehensiveness of the appearance defect detection of the rubber strips.

[0071] Furthermore, according to the different shooting angles between the shooting component and the rubber strip, the system can accurately analyze the defect contour of the rubber strip according to the angle of the shooting image, thereby accurately detecting the appearance defect type of the rubber strip.

[0072] Specifically, image parameters are extracted from the characteristic contour image, and the image parameters include shape and size data of the characteristic contour. According to the contour of the characteristic contour image, the defect type of the characteristic contour image can be preliminarily screened. Combined with the size data of the characteristic contour image, the detection model can further screen the defect type corresponding to the characteristic contour image to improve the accuracy of characteristic contour image detection.

[0073] S14: determining a plurality of groups of defect comparison data according to the shooting parameters of the characteristic contour image, and analyzing the similarity between the characteristic contour image and the defect comparison data according to the image parameters.

[0074] Extracting defect image data corresponding to the shooting parameters from a database of a detection model according to the shooting parameters of the characteristic contour image;

[0075] The defect image data is divided into several groups of defect comparison data according to the type of processing defects of the rubber strip appearance.

[0076] Screening a number of groups of defect comparison data according to the shape data of the characteristic contour image to obtain a number of groups of similar defect comparison data;

[0077] According to the size data of the characteristic contour image, a structural similarity analysis is performed on several groups of similar defect comparison data, and the analysis results are output.

[0078] The calculation formula for the structural similarity analysis is:

[0079]

[0080] Among them, SSIM(X,Y) is the similarity function, μ x is the average value of X, μ y is the average value of Y, δ x 2 is the variance of X, δ y 2 is the variance of Y, δ xy is the covariance of X and Y, and C1 and C2 are constants.

[0081] Specifically, the structural similarity analysis is mainly based on the comparison of the brightness, contrast and structure of the feature contour image and the defect comparison data, so as to comprehensively analyze the similarity between the feature contour image and the defect comparison data.

[0082] S15: Determine defect comparison data with the highest matching degree with the characteristic contour image according to the similarity, and output the defect type of the rubber strip.

[0083] By performing similarity analysis on several groups of defect control data and characteristic contour images, the similarity data between the characteristic contour image and each group of defect control data are output, and they are sorted from high to low according to the similarity, and the defect type corresponding to the defect control data with the highest similarity is marked as the defect type corresponding to the characteristic contour image.

[0084] Specifically, the detection method further includes:

[0085] After confirming the defect type of the characteristic contour image, data enhancement is performed on the characteristic contour to obtain a data set of the characteristic contour. The characteristic contour image of the confirmed defect type is subjected to data enhancement operations such as geometric transformation, random flipping, segmentation and recombination to obtain a number of image data, thereby obtaining a data set of the characteristic contour.

[0086] The data set is divided into a training set and a test set by random division, the defect detection model is trained based on the training set, and the model training result is verified by the test set.

[0087] The training set is input into the defect detection model, and the defect detection model is trained by the training set to improve the defect detection model. During the actual shooting and detection process of the rubber strips, the defect detection model can be trained and corrected according to each batch of rubber strips, so that the defect detection model can accurately detect and identify the defect type of the rubber strips, thereby improving the accuracy of the appearance detection of the rubber strips.

[0088] In addition, by generating multiple types of inspection images, subsequent defect analysis can be performed from multiple angles, greatly improving the comprehensiveness and accuracy of inspection. For example, some defects may not be obvious in grayscale images, but can be clearly displayed in edge detection images. This multi-dimensional analysis method significantly reduces the missed detection rate and improves the reliability of inspection.

[0089] Embodiment 2:

[0090] Figure 2 The figure shows a connection diagram of a device for detecting defects in the appearance of a rubber strip according to an embodiment of the present invention. Figure 3 The schematic diagram of the structure of the device for detecting the appearance processing defects of the rubber strip in the embodiment of the present invention is shown. The detection device is used to execute the above-mentioned method for detecting the appearance processing defects of the rubber strip. The detection device comprises:

[0091] The transmission mechanism 10 includes a first transmission belt and a second transmission belt arranged in parallel, and a transparent adhesive film arranged between the first transmission belt and the second transmission belt, and a light detection area is arranged in the middle of the transmission mechanism;

[0092] Lighting mechanism 20: arranged in the light detection area, used for lighting the rubber strip;

[0093] Photographing component 30: used for photographing the rubber strip after light treatment to obtain a detection image of the rubber strip;

[0094] Main control system 40: used to process the inspection image of the rubber strip and obtain the appearance processing defects of the rubber strip.

[0095] Specifically, an inclined transition section is provided in the middle of the transmission mechanism, and the transition section forms the light detection area. The transmission mechanism is divided into an upstream part and a downstream part based on the transition section. The rubber strip workpiece is input into the transmission mechanism from the upstream part, and the rubber strip workpiece is transported based on the transmission mechanism, and the rubber strip workpiece is polished in the transition section so as to perform defect detection on the rubber strip workpiece.

[0096] The lighting mechanism is arranged in the lighting detection area to illuminate the rubber strip. The lighting mechanism can adopt a variety of light sources, such as LED lights, halogen lamps or fluorescent lamps. In order to obtain the best detection effect, the lighting mechanism can be arranged below the transparent film of the transmission mechanism, and the light output direction is toward the transparent film. In this way, the rubber strip can be evenly illuminated through the transparent film, avoiding the reflection or shadow problems that may be caused by the direct light source. In addition, the lighting mechanism can be designed to output flexible light, and this soft light can better highlight the subtle defects on the surface of the rubber strip.

[0097] The shooting component is used to shoot the rubber strip after lighting treatment and obtain the inspection image of the rubber strip. The shooting component can include one or more high-resolution cameras, which can shoot the rubber strip from different angles to obtain comprehensive image information. The position and angle of the shooting component need to be precisely adjusted to ensure that all details on the surface of the rubber strip can be captured.

[0098] The main control system is used to process the inspection image of the rubber strip and obtain the appearance processing defects of the rubber strip. The main control system can use a high-performance computer or a dedicated image processor, equipped with advanced image processing algorithms and machine learning models. These algorithms can perform preprocessing, feature extraction, defect recognition and classification on the captured images.

[0099] In actual applications, the rubber strip is transported to the light detection area through the first conveyor belt and the second conveyor belt. In this area, the light mechanism illuminates the rubber strip to make the defects on the surface of the rubber strip more obvious. The camera component then takes a picture of the rubber strip after the light treatment to obtain high-quality detection images. These images are transmitted to the main control system, which analyzes these images through complex image processing algorithms to identify various defects on the surface of the rubber strip, such as scratches, bubbles, impurities, etc.

[0100] When the rubber strip is facing the camera assembly, that is, the rubber strip is perpendicular to the shooting direction of the camera assembly, the forward shooting method can improve the accuracy of detecting appearance defects such as bubbles, impurities, oil rings, glue, flow marks, cracks, uneven sandblasting, etc. on the rubber strip.

[0101] When the rubber strip and the shooting direction of the camera assembly form an acute angle or an obtuse angle, the accuracy of detecting appearance defects such as crushing, creases, and deformation of the rubber strip can be improved.

[0102] Based on the rubber strips in different postures, the camera component can improve the defect detection of different observation surfaces and different positions of the rubber strips, thereby improving the comprehensiveness of the appearance defect detection of the rubber strips.

[0103] Furthermore, according to the different shooting angles between the shooting component and the rubber strip, the system can accurately analyze the defect contour of the rubber strip according to the angle of the shooting image, thereby accurately detecting the appearance defect type of the rubber strip.

[0104] This design realizes the automation and standardization of the inspection of the appearance processing defects of the rubber strip. The design of the transmission mechanism ensures the stability and continuity of the rubber strip during the inspection process. The use of the lighting mechanism improves the visibility of the defects and helps to improve the accuracy of the inspection. The cooperation of the shooting component and the main control system realizes the automatic identification and analysis of defects, greatly improving the inspection efficiency.

[0105] An inclined transition section is set in the middle of the transmission mechanism as the light detection area. This design can change the transmission angle of the rubber strip, so that the rubber strip presents different surfaces in the light detection area, which is conducive to comprehensive detection of defects on the surface of the rubber strip. The inclined design can also increase the length of the light detection area, extend the residence time of the rubber strip in the detection area, and improve the adequacy and accuracy of the detection. The setting of the transition section creates a height difference between the light detection area and other transmission areas, which helps to control the range of light exposure and reduce the interference of ambient light on the detection. The inclined arrangement can keep a certain tension on the rubber strip in the detection area to avoid detection errors caused by loose rubber strips.

[0106] Specifically, the inclined transition section can be set within an angle range of 15° to 45°, which can provide sufficient inclination to achieve the above effect while ensuring smooth transmission of the rubber strip. The length of the transition section can be adjusted according to the width of the rubber strip and the detection requirements, and can usually be set to 1.5 to 3 times the width of the rubber strip. Such a design can ensure that the rubber strip has sufficient exposure area and residence time in the light detection area.

[0107] The transition section may be provided with a plurality of undulating positions, and based on the undulating positions, the conveying posture of the rubber strip is changed when the rubber strip passes through the transition section, so that the camera assembly can obtain the rubber strip in different posture states.

[0108] According to the photographed images of the rubber strip in different positions and postures, the accuracy of identifying the type of rubber strip defects is improved.

[0109] Compared with the traditional planar transmission structure, the design of the embodiment of the present invention forms a light detection area through the inclined transition section, which significantly improves the comprehensiveness and accuracy of the detection. The traditional planar transmission structure can only fully detect one side of the rubber strip, while the inclined design of the embodiment of the present invention allows the rubber strip to present a surface at multiple angles during the transmission process, so that the lighting and camera components can capture more surface details.

[0110] In addition, the design of the embodiment of the present invention also solves the problem of uneven illumination in traditional detection devices. By arranging the light source in an inclined manner, the light source can be more evenly illuminated to the surface of the rubber strip, reducing the generation of shadows and blind spots. This not only improves the accuracy of detection, but also reduces the probability of missed detection and false detection.

[0111] This technical solution achieves effective illumination of the adhesive strip by placing the illumination mechanism below the transparent adhesive film of the transmission mechanism and directing the light emission direction of the illumination mechanism toward the transparent adhesive film. This arrangement has multiple advantages, including protecting the illumination mechanism, improving the lighting effect, making full use of the light transmittance of the transparent adhesive film, and optimizing the spatial layout.

[0112] Specifically, the illumination mechanism can be implemented in a variety of forms. For example, light sources such as LED light strips, fluorescent tubes, or halogen lamps can be used. These light sources can be evenly distributed along the length of the transmission mechanism to ensure uniform illumination of the entire detection area. The installation position of the illumination mechanism can be achieved by an adjustable bracket, so that it can be fine-tuned according to the thickness of the transparent film.

[0113] This arrangement forms a coordinated whole with the transmission mechanism and the shooting component. As an intermediate medium, the transparent film not only transmits the rubber strip, but also scatters and homogenizes the light. This optical path design reduces the strong light reflection or shadow problems that may be caused by direct irradiation, providing more ideal lighting conditions for the shooting component. At the same time, because the lighting mechanism is located below the rubber strip, the space above is released, which makes it possible to flexibly arrange the shooting component, thereby realizing multi-angle and multi-directional image acquisition.

[0114] In practical applications, the light intensity of the illumination mechanism can be adjusted according to the material and thickness of the adhesive strip. For example, for thicker or less transparent adhesive strips, the light intensity can be appropriately increased; for thinner or more transparent adhesive strips, the light intensity can be reduced to avoid overexposure. This flexible illumination control can be automatically adjusted by the main control system, further improving the adaptability and accuracy of detection.

[0115] This design of the embodiment of the present invention not only solves the lighting problem, but also takes into account the protection of the equipment and the use of space, reflecting the comprehensiveness and practicality of the design. Through the optimized lighting mechanism setting, the detection device can obtain a clearer and more uniform image of the rubber strip, thereby improving the accuracy and efficiency of the appearance processing defect detection.

[0116] The illumination mechanism outputs flexible light outwards. Flexible light is more uniform and soft, which can reduce the reflection and glare of strong light on the surface of the rubber strip, thereby improving the quality of the inspection image. By using flexible light, subtle defects on the surface of the rubber strip, such as scratches and bubbles, can be better highlighted. Flexible light can evenly illuminate the surface of the rubber strip, reduce shadows and highlight areas, and make defect features more clearly visible. This helps in the subsequent image processing and defect recognition process, and improves the accuracy and reliability of detection. At the same time, flexible light has little effect on the rubber strip itself, and will not cause deformation or damage to the rubber strip due to strong light exposure. This lighting method can both meet the inspection needs and protect the integrity of the rubber strip.

[0117] The illumination mechanism in the embodiment of the present invention can achieve the output of flexible light in a variety of ways. For example, an LED soft light plate can be used, which is composed of a plurality of small LED lamp beads, and the light is evenly dispersed by a diffuse reflection plate or a diffusion plate. Another way is to use a fiber optic lighting system, in which the light generated by the light source is transmitted through the optical fiber and scattered at the end to achieve a soft and uniform lighting effect. Reflective lighting can also be used, using a reflector or a reflective plate to reflect and scatter the light emitted by the light source, thereby producing soft light.

[0118] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0119] In addition, the above provides a detailed introduction to the embodiments of the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for detecting defects in the appearance of a rubber strip, characterized in that: The detection method comprises: Setting a plurality of photographing points on the light detection area, and obtaining a plurality of detection images of the adhesive strip at the plurality of photographing points; Compare the detection image with the standard image to obtain the characteristic contour of the detection image; Input the feature contour image into the defect detection model, and extract the shooting parameters and image parameters of the feature contour image; Determine a plurality of groups of defect comparison data according to shooting parameters of the characteristic contour image, and analyze similarity between the characteristic contour image and the defect comparison data according to the image parameters; Defect comparison data with the highest matching degree with the characteristic contour image is determined according to the similarity, and the defect type of the rubber strip is output.

2. The method for detecting the appearance processing defects of the rubber strip according to claim 1, characterized in that: The step of setting a plurality of photographing points on the light detection area and obtaining a plurality of detection images of the adhesive strip at the plurality of photographing points comprises: A plurality of rubber strips are transported to a light detection area based on a conveyor belt, and the rubber strips are illuminated based on a light mechanism; A plurality of photographing points are set on the light detection area, and the rubber strip after the light treatment is photographed and detected by a camera component to obtain a plurality of detection images of the rubber strip at the plurality of photographing points.

3. The method for detecting the appearance processing defects of the rubber strip according to claim 1, characterized in that: The step of comparing the detection image with the standard image to obtain the characteristic contour of the detection image comprises: Preprocess the detection image and convert it into a pixel image; The standard image of the rubber strip is extracted from the database, and after the standard image is converted into a pixel map, the pixel map of the standard image is subtracted from the pixel map of the detection image to extract the feature contour in the detection image.

4. The method for detecting the appearance processing defects of the rubber strip according to claim 1, characterized in that: The step of inputting the characteristic contour image into the defect detection model and extracting the shooting parameters and image parameters of the characteristic contour image comprises: Acquire shooting parameters of the characteristic contour image according to the detection image corresponding to the characteristic contour image, wherein the shooting parameters include a shooting angle and a shooting position; Image parameters are extracted from the characteristic contour image, wherein the image parameters include shape and size data of the characteristic contour.

5. The method for detecting the appearance processing defects of the rubber strip according to claim 4, characterized in that: Determining a plurality of groups of defect comparison data according to the shooting parameters of the characteristic contour image includes: Extracting defect image data corresponding to the shooting parameters from a database of a detection model according to the shooting parameters of the characteristic contour image; The defect image data is divided into several groups of defect comparison data according to the type of processing defects of the rubber strip appearance.

6. The method for detecting the appearance processing defects of the rubber strip according to claim 5, characterized in that: The analyzing the similarity between the feature contour image and the defect comparison data according to the image parameters comprises: Screening a number of groups of defect comparison data according to the shape data of the characteristic contour image to obtain a number of groups of similar defect comparison data; According to the size data of the characteristic contour image, a structural similarity analysis is performed on several groups of similar defect comparison data, and the analysis results are output.

7. The method for detecting the appearance processing defects of the rubber strip according to claim 6, characterized in that: The calculation formula for the structural similarity analysis is: Among them, SSIM(X,Y) is the similarity function, μ x is the average value of X, μ y is the average value of Y, δ x 2 is the variance of X, δ y 2 is the variance of Y, δ xy is the covariance of X and Y, and C1 and C2 are constants.

8. The method for detecting the appearance processing defects of rubber strips according to claim 1, characterized in that: The detection method further comprises: After confirming the defect type of the feature contour image, data enhancement is performed on the feature contour to obtain a data set of the feature contour; The data set is divided into a training set and a test set, the defect detection model is trained based on the training set, and the model training result is verified by the test set.

9. A device for detecting defects in the appearance of rubber strips, characterized in that: The detection device is used to perform the method for detecting the appearance processing defects of the rubber strip as claimed in any one of claims 1 to 8; The detection device comprises: The transmission mechanism comprises a first transmission belt and a second transmission belt arranged in parallel, and a transparent adhesive film arranged between the first transmission belt and the second transmission belt, and a light detection area is arranged in the middle of the transmission mechanism; Illumination mechanism: arranged in the illumination detection area, used for performing illumination treatment on the rubber strip; Shooting component: used to shoot the rubber strip after lighting treatment and obtain the detection image of the rubber strip; Main control system: used to process the inspection image of the rubber strip and obtain the appearance processing defects of the rubber strip.

10. The device for detecting defects in the appearance of rubber strips according to claim 9, characterized in that: The middle part of the transmission mechanism is provided with a plurality of wave segments at undulating positions, and the wave segments form the light detection area.

Images