A method and system for detecting surface defects in foamed materials

Abstract

This invention relates to the field of defect detection technology, specifically to a method and system for detecting surface defects in foamed materials. The method includes the following steps: using a camera sensor to scan the area of ​​the foamed material to be tested to acquire an RGB color image; performing image augmentation processing on the acquired RGB color image to improve its clarity; coarsely registering the pre-processed RGB color image with a defect-free standard image to obtain images with and without obvious defects; and finely registering the image without obvious defects with the defect-free standard image to identify images with minor defects. If minor defects are found in the image without obvious defects, it is marked as an image with minor defects. Through the combination of these steps, this invention enables the detection of whether defects exist on the surface of foamed materials, and can detect whether the defects are obvious or minor, as well as the number of defects, thus improving detection efficiency and accuracy.

Description

Technical Field

[0001] This invention relates to the technical field of surface defect detection in foamed materials, and more specifically, to a method and system for surface defect detection in foamed materials. Background Technology

[0002] Foamed materials are materials with unique properties and are widely used in many fields. Foamed materials refer to porous materials formed by generating a large number of air bubbles inside the material through physical or chemical methods. These air bubbles are evenly distributed in the material matrix, giving the material characteristics such as low density, high specific strength, and good thermal and sound insulation properties.

[0003] The performance of foamed materials (such as thermal insulation, cushioning, and mechanical properties) is closely related to their microstructure and overall uniformity. Defects (such as uneven pores and cracks) can compromise the structural integrity of the material, leading to unstable performance. For example, in the field of building insulation, the presence of pores or cracks in foamed materials will reduce their thermal insulation performance and affect the energy efficiency of buildings. In the packaging industry, if foamed materials used for cushioning protection have cracks or localized density abnormalities, they may not be able to effectively absorb impact energy, resulting in damage to the packaged items. Therefore, it is necessary to inspect the surface of foamed materials for defects after production.

[0004] For example, Chinese Patent Publication No. CN108230321A discloses a defect detection method and device. This comparative patent compares the gray value of the pixel in the variance image of the region with a specified threshold. If there is a pixel whose gray value exceeds the specified threshold, the location of the pixel can be determined as a defect area, and the defect area can be marked, thereby realizing automated detection of defects on the surface of the cover glass.

[0005] Although the aforementioned comparative patent can determine whether the area to be detected is a defective area by comparing the gray value of the area to be detected with a specified threshold, the detection method provided has low detection accuracy and incomplete detection results. Some minor defects that are not obvious do not affect the normal sales and use of certain products, leading to the sale of products with defects. Therefore, it is necessary to improve it. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for detecting surface defects in foamed materials. If an image without obvious defects is consistent with a standard image without defects, then the image without obvious defects does not contain any inconspicuous defects, i.e., the image without obvious defects does not contain any defects. This invention aims to solve the problems in the prior art.

[0007] This invention is implemented as follows: a method for detecting surface defects in foamed materials, applied to surface defect detection equipment, specifically includes the following steps:

[0008] S101. Display the editing control on the main interface, determine whether to obtain the summarization operation of the editing control, and if the surface defect detection device obtains the summarization operation of the editing control, then use the camera sensor to scan the area of ​​the foamed material to be tested to obtain a set of RGB color images.

[0009] S102. Perform image augmentation and adaptation preprocessing on the acquired RGB color image set to improve the clarity of the RGB color images;

[0010] S103. Perform a first recognition registration between the RGB color image after image augmentation and adaptation preprocessing and the system's preset defect-free standard image. If there is no obvious difference between the RGB color image and the defect-free standard image, mark the RGB color image as an image without obvious defects. Perform a second recognition registration between the image without obvious defects and the defect-free standard image. If the image without obvious defects is consistent with the defect-free standard image, then the image without obvious defects does not have any inconspicuous defects, i.e., the image without obvious defects does not have any defects. If the image without obvious defects is inconsistent with the defect-free standard image, then the image without obvious defects has inconspicuous defects, and mark the image without obvious defects as an image with inconspicuous defects.

[0011] S104. If there is a significant difference between the RGB color image and the defect-free standard image, the RGB color image shall be marked as an image with significant defects.

[0012] S105. Perform a second recognition registration between the image with obvious defects and the standard image without defects, identify the defect feature points in the image with obvious defects and the standard image without defects, and mark the identified defect feature points in the image with obvious defects and the standard image without defects according to the inductive operation of the obtained editing control. After the marking is completed, a defect set is formed and output to the terminal.

[0013] Further, in S101, if the surface defect detection device acquires the summarization operation of the editing control, and then uses the camera sensor to scan the area of ​​the foamed material to be tested to acquire a set of RGB color images, it includes:

[0014] The summary request for editing controls is sent by a preset network receiving terminal, and the summary request for editing controls is used to request the establishment of a connection with the surface defect detection device;

[0015] If the summarization request of the editing control sent by the terminal contains a summarization set of editing controls, then the terminal is connected according to the summarization request of the editing control. After the connection is completed, the RGB color image set is obtained by scanning the area of ​​the foaming material to be tested using a camera sensor.

[0016] The camera sensor can be any type of video camera or infrared camera.

[0017] Furthermore, in S102, the acquired RGB color image set is subjected to image augmentation and adaptation preprocessing. The image augmentation and adaptation preprocessing features include hue, saturation, brightness, contrast changes, image flipping, rotation, image displacement, and center segmentation.

[0018] Further, in S103, the first identification registration includes:

[0019] Obtain recognition models for multiple defect-free standard images, and sort the recognition models for the multiple defect-free standard images in order;

[0020] The preprocessed RGB color image is obtained and input into the recognition model of the first ranked defect-free standard image for recognition. If no imperfection or damage is directly identified, the RGB color image is a defect-free image.

[0021] If imperfections or damage are directly identified, the RGB color image is considered to be an image with obvious defects.

[0022] Imperfections or damage include, but are not limited to, scratches, dents, cracks, corrosion, uneven coating, and holes.

[0023] Further, in S103, a second identification registration is performed between the image without obvious defects and the standard image without defects. The fine registration of the image without obvious defects includes the following steps:

[0024] S1031. Divide the image without obvious defects and the standard image without defects into multiple small image regions equally. Enlarge and brighten all small image regions on the image without obvious defects and the standard image without defects. Then compare all small image regions of the two one by one. If all small image regions of the two are consistent, then the image without obvious defects does not have any inconspicuous defects.

[0025] S1032. If there are inconsistent areas in all small image regions of the two, then the image without obvious defects has inconspicuous defects.

[0026] Furthermore, S1032 also includes the following steps:

[0027] S10321. Locate all inconsistent small image regions;

[0028] S10322. The small image region that has been located is partially masked by the generated background masking region in order to better focus on the region with inconspicuous defects and generate a new local positioning region.

[0029] S10323. Enlarge and crop the new local positioning area to capture more detailed information, making the inconspicuous features in the small image area more obvious, so as to extract the inconspicuous defect feature information more effectively, so as to compare the inconsistent areas on the image without obvious defects and the standard image without defects.

[0030] Furthermore, in S1032, inconspicuous defects include, but are not limited to, microcracks and minor deformations or misalignments.

[0031] Furthermore, in S10323, the new localized region is enlarged and cropped to capture more detailed information, making inconspicuous features in small image regions more apparent, including:

[0032] The localized small image region is obtained, the edge contour of the small image region is identified and the edge contour feature is generated, and a background occlusion region is generated outside the edge contour feature to shield part of the background of the small image region.

[0033] Identify and mark areas with high chromatic aberration within the edge contour features. Then, magnify and crop the marked areas with high chromatic aberration to capture more detailed information.

[0034] Compared with the prior art, the method and system for detecting surface defects in foamed materials provided by the present invention have the following advantages:

[0035] 1. This invention can detect whether there are obvious defects on the surface of foaming materials by performing a first identification registration between the RGB color image after image augmentation and adaptation preprocessing and the defect-free standard image.

[0036] If there is no significant difference between the RGB color image and the defect-free standard image, the RGB color image is marked as an image without significant defects, meaning that the surface of the foamed material being inspected has no significant defects. If there is a significant difference between the RGB color image and the defect-free standard image, the RGB color image is marked as an image with significant defects, meaning that the surface of the foamed material being inspected has significant defects, i.e., there are defects.

[0037] 2. By performing a second recognition registration between an image without obvious defects and a standard image without defects, this invention can detect whether an image without obvious defects exists.

[0038] If an image without obvious defects is consistent with a standard image without defects, then the image without obvious defects does not contain any inconspicuous defects, i.e., the image without obvious defects does not contain any defects. If an image without obvious defects is inconsistent with a standard image without defects, then the image without obvious defects contains inconspicuous defects, and the image without obvious defects is marked as an image with inconspicuous defects.

[0039] A surface defect detection system for foamed materials, used to perform the above-described surface defect detection method for foamed materials, the surface defect detection system comprising:

[0040] The image acquisition unit is used to acquire RGB color images of the area of ​​the foam material to be tested;

[0041] Image augmentation processing unit, used to improve the clarity of the RGB color image;

[0042] The coarse registration unit is used to detect whether there are obvious defects in the RGB color image;

[0043] The fine registration unit is used to detect subtle defects in RGB color images.

[0044] Specifically, the fine registration unit includes a positioning module, used to locate all inconsistent small image regions;

[0045] The background masking module masks a portion of the background in a small image area by generating a background masking region, thus preserving the features of inconspicuous defect areas.

[0046] The zoom and crop module is used to zoom in and crop small image areas with masked backgrounds, so as to extract less obvious defect feature information more effectively. Attached Figure Description

[0047] Figure 1 This is a schematic flowchart of a method for detecting surface defects in foamed materials proposed in this invention.

[0048] Figure 2 This is a flowchart illustrating the process of the surface defect detection device acquiring the summary operation of the editing control in a method for detecting surface defects in foamed materials proposed in this invention.

[0049] Figure 3 This is a schematic diagram of the structure of a foam material surface defect detection system proposed in this invention;

[0050] Figure 4 This is a schematic diagram of the structure of a fine registration unit in a foam material surface defect detection system proposed in this invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0052] The implementation of the present invention will be described in detail below with reference to specific embodiments.

[0053] In the accompanying drawings of this embodiment, the same or similar reference numerals correspond to the same or similar components. In the description of this invention, it should be understood that if terms such as "upper," "lower," "left," and "right" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting this invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0054] The main types of surface defects in foamed materials include the following:

[0055] 1. Rough edges: Rough edges refer to the unevenness, irregularity, or burrs on the surface of foamed products. This defect can affect the product's appearance and performance.

[0056] 2. Air bubbles: Air bubbles are cavities filled with gas that form inside foamed products, affecting the product's sealing performance, mechanical properties, and appearance. The formation of air bubbles is related to factors such as moisture or impurities in the raw materials, unstable mold temperature or failure to reach a suitable temperature, improper temperature and pressure control during the foaming process, and unreasonable foaming technology.12

[0057] 3. Cells: Cells appearing on the surface of foamed boards are a sign of locally low melt density. They generally appear as either streaks or round holes. The occurrence of cells is related to raw material factors such as resin, foaming agent, regulator, calcium carbonate, and stabilizer, as well as molds and processes.

[0058] 4. Segregation: Segregation refers to uneven foam distribution, with unreacted raw materials interspersed within. The foam will feel damp and sticky to the touch. Segregation is caused by uneven mixing of P and I components during the reaction, lack of pre-mixing or insufficient mixing of raw materials, incorrect raw material ratio settings, abnormal equipment flow metering, and abnormal mixing pressure.

[0059] 5. Deficient Foam: Deficient foam refers to a product that appears without obvious defects upon visual inspection, but feels noticeably collapsed to the touch. Cutting open the surface reveals coarse, fibrous foam cells. Deficient foam is caused by various factors, including insufficient internal pressure, a sudden drop in internal pressure due to material overflow, mismatch between mold air cushion deflation time and material / mold temperature leading to insufficient foam flowability due to product design, insufficient casting material, and expired or unevenly mixed foaming material.

[0060] To detect surface defects in foamed materials, refer to Figures 1-2As shown, a method for detecting surface defects in foamed materials, applied to surface defect detection equipment, specifically includes the following steps:

[0061] S101. Display the editing control on the main interface and determine whether to obtain the summary operation of the editing control. If the surface defect detection device obtains the summary operation of the editing control, then use the camera sensor to scan the area of ​​the foamed material to be tested to obtain a set of RGB color images.

[0062] Specifically, if the surface defect detection equipment acquires the summarization operation of the editing control, and then uses the camera sensor to scan the area of ​​the foamed material to be tested to obtain a set of RGB color images, including:

[0063] The summary request for the editing control is sent through the preset network receiving terminal. The summary request for the editing control is used to request the establishment of a connection with the surface defect detection equipment.

[0064] If the summary request for the editing control sent by the terminal contains a summary set of editing controls, then the terminal is connected according to the summary request for the editing control. After the connection is completed, the camera sensor is used to scan the area of ​​the foaming material to be tested to obtain a set of RGB color images.

[0065] The camera sensor can be either a video camera or an infrared camera;

[0066] S102. Perform image augmentation and adaptation preprocessing on the acquired RGB color image set to improve the clarity of the RGB color images;

[0067] The acquired RGB color image set is subjected to image augmentation and adaptation preprocessing. The image augmentation and adaptation preprocessing features include hue, saturation, brightness, contrast changes, image flipping, rotation, image displacement, and center segmentation.

[0068] S103. Perform a first recognition registration between the RGB color image after image augmentation and adaptation preprocessing and the system's preset defect-free standard image. If there is no obvious difference between the RGB color image and the defect-free standard image, mark the RGB color image as an image without obvious defects. Perform a second recognition registration between the image without obvious defects and the defect-free standard image. If the image without obvious defects is consistent with the defect-free standard image, then the image without obvious defects does not have any inconspicuous defects, i.e., the image without obvious defects does not have any defects. If the image without obvious defects is inconsistent with the defect-free standard image, then the image without obvious defects has inconspicuous defects, and mark the image without obvious defects as an image with inconspicuous defects.

[0069] Specifically, by performing a first identification registration between the preprocessed RGB color image after image augmentation and adaptation and a defect-free standard image, it is possible to detect whether there are obvious defects on the surface of the foaming material.

[0070] If there is no significant difference between the RGB color image and the defect-free standard image, the RGB color image is marked as an image without significant defects, meaning that the surface of the foamed material being inspected has no significant defects; if there is a significant difference between the RGB color image and the defect-free standard image, the RGB color image is marked as an image with significant defects, meaning that the surface of the foamed material being inspected has significant defects, i.e., there are defects.

[0071] The process involves a second identification registration between an image without obvious defects and a standard image without defects. The fine registration of the image without obvious defects includes the following steps:

[0072] S1031. Divide the image without obvious defects and the standard image without defects into multiple small image regions equally. Enlarge and brighten all small image regions on the image without obvious defects and the standard image without defects. Then compare all small image regions of the two one by one. If all small image regions of the two are consistent, then the image without obvious defects does not have any inconspicuous defects.

[0073] S1032. If there are inconsistent areas in all small image regions of the two, then the image without obvious defects has inconspicuous defects.

[0074] S104. If there is a significant difference between the RGB color image and the defect-free standard image, the RGB color image shall be marked as an image with significant defects.

[0075] S105. Perform a second recognition registration between the image with obvious defects and the standard image without defects to identify defect feature points in both images. Based on the inductive operation of the acquired editing controls, mark the identified defect feature points in both images. After marking, a defect set is formed and output to the terminal. This invention, by performing a second recognition registration between the image without obvious defects and the standard image without defects, can detect whether an image without obvious defects exists or not. If the image without obvious defects is consistent with the standard image without defects, then the image without obvious defects does not have any notable defects, i.e., the image without obvious defects does not have any defects. If the image without obvious defects is inconsistent with the standard image without defects, then the image without obvious defects contains notable defects, and the image without obvious defects is marked as an image with notable defects.

[0076] In S103 of this embodiment, the first identification registration includes:

[0077] Obtain recognition models for multiple defect-free standard images, and sort the recognition models of multiple defect-free standard images in order;

[0078] The preprocessed RGB color image is obtained and input into the recognition model of the first-ranked defect-free standard image. If no imperfections or damages are directly identified, the RGB color image is considered to be a defect-free image.

[0079] If imperfections or damage are directly identified, the RGB color image is considered to be an image with obvious defects.

[0080] Imperfections or damage include, but are not limited to, scratches, dents, cracks, corrosion, uneven coating, and holes.

[0081] In S1032 of this embodiment, the following step is also included:

[0082] S10321. Locate all inconsistent small image regions;

[0083] S10322. The small image region that has been located is partially masked by the generated background masking region in order to better focus on the region with inconspicuous defects and generate a new local positioning region.

[0084] S10323. Enlarge and crop the new local positioning area to capture more detailed information, making the inconspicuous features in the small image area more obvious, so as to extract the inconspicuous defect feature information more effectively, so as to compare the inconsistent areas on the image without obvious defects and the standard image without defects.

[0085] In S1032, inconspicuous defects include, but are not limited to, microcracks and minor deformations or misalignments.

[0086] In S10323, the new localized region is enlarged and cropped to capture more detailed information, making inconspicuous features in small image regions more apparent, including:

[0087] The localized small image region is obtained, the edge contour of the small image region is identified and the edge contour feature is generated, and a background occlusion region is generated outside the edge contour feature to shield part of the background of the small image region.

[0088] Identify and mark areas with high chromatic aberration within the edge contour features. Then, magnify and crop the marked areas with high chromatic aberration to capture more detailed information.

[0089] Reference Figures 3-4As shown, a surface defect detection system for foamed materials is used to perform the above-mentioned surface defect detection method for foamed materials. The surface defect detection system includes: an image acquisition unit for acquiring RGB color images of the area of ​​the foamed material to be tested; an image augmentation processing unit for improving the clarity of the RGB color images; a coarse registration unit for detecting whether there are obvious defects in the RGB color images; and a fine registration unit for detecting whether there are inconspicuous defects in the RGB color images. This invention can detect whether there are obvious defects on the surface of the foamed material by performing a first identification registration between the RGB color images after image augmentation and adaptation preprocessing and a defect-free standard image. If there is no obvious difference between the RGB color images and the defect-free standard image, the RGB color images are marked as images without obvious defects, that is, the surface of the inspected foamed material has no obvious defects. If there is an obvious difference between the RGB color images and the defect-free standard image, the RGB color images are marked as images with obvious defects, that is, the surface of the inspected foamed material has obvious defects, i.e., defects exist.

[0090] The fine registration unit includes a positioning module for locating all inconsistent small image regions; a background masking module for masking a portion of the background of the small image regions by generating a background masking area, thus preserving the features of inconspicuous defect regions; and a magnification and cropping module for magnifying and cropping the small image regions with masked backgrounds to more effectively extract inconspicuous defect feature information. By performing a second recognition registration between the image without obvious defects and the standard image without defects, it is possible to detect whether the image without obvious defects contains inconspicuous defects. If the image without obvious defects is consistent with the standard image without defects, then the image without obvious defects does not contain inconspicuous defects, i.e., the image without obvious defects does not contain defects. If the image without obvious defects is inconsistent with the standard image without defects, then the image without obvious defects contains inconspicuous defects, and the image without obvious defects is marked as an image with inconspicuous defects.

[0091] In this embodiment, the entire operation process can be controlled by a computer, along with a PLC, to achieve automated operation control. In each operation stage, sensors can be set up to provide signal feedback and ensure that the steps are performed sequentially. These are all conventional knowledge in current automation control, and will not be elaborated on in this embodiment.

[0092] 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, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting surface defects in foamed materials, characterized in that: Applied to surface defect detection equipment, the specific steps include: S101. Display the editing control on the main interface, determine whether to obtain the summarization operation of the editing control, and if the surface defect detection device obtains the summarization operation of the editing control, then use the camera sensor to scan the area of ​​the foamed material to be tested to obtain a set of RGB color images. S102. Perform image augmentation and adaptation preprocessing on the acquired RGB color image set to improve the clarity of the RGB color images; S103. Perform a first recognition registration between the RGB color image after image augmentation and adaptation preprocessing and the system's preset defect-free standard image. If there is no obvious difference between the RGB color image and the defect-free standard image, mark the RGB color image as an image without obvious defects. Perform a second recognition registration between the image without obvious defects and the defect-free standard image. If the image without obvious defects is consistent with the defect-free standard image, then the image without obvious defects does not have any inconspicuous defects, i.e., the image without obvious defects does not have any defects. If the image without obvious defects is inconsistent with the defect-free standard image, then the image without obvious defects has inconspicuous defects, and mark the image without obvious defects as an image with inconspicuous defects. In S103, the first identification registration includes: Obtain recognition models for multiple defect-free standard images, and sort the recognition models for the multiple defect-free standard images in order; The preprocessed RGB color image is obtained and input into the recognition model of the first ranked defect-free standard image for recognition. If no imperfection or damage is directly identified, the RGB color image is a defect-free image. If imperfections or damage are directly identified, the RGB color image is considered to be an image with obvious defects. Imperfections or damage include, but are not limited to, scratches, dents, cracks, corrosion, uneven coating, and holes; A second identification registration is performed between the image without obvious defects and the standard image without defects. The fine registration of the image without obvious defects includes the following steps: S1031. Divide the image without obvious defects and the standard image without defects into multiple small image regions equally. Enlarge and brighten all small image regions on the image without obvious defects and the standard image without defects. Then compare all small image regions of the two one by one. If all small image regions of the two are consistent, then the image without obvious defects does not have any inconspicuous defects. S1032. If there are inconsistent areas in all small image regions of the two, then the image without obvious defects has inconspicuous defects. S1032 also includes the following steps: S10321. Locate all inconsistent small image regions; S10322. The small image region that has been located is partially masked by the generated background masking region in order to better focus on the region with inconspicuous defects and generate a new local positioning region. S10323. Enlarge and crop the new local positioning area to capture more detailed information, making the inconspicuous features in the small image area more obvious, so as to extract the inconspicuous defect feature information more effectively, so as to compare the inconsistent areas on the image without obvious defects and the standard image without defects. S104. If there is a significant difference between the RGB color image and the defect-free standard image, the RGB color image shall be marked as an image with significant defects. S105. Perform a second recognition registration between the image with obvious defects and the standard image without defects, identify the defect feature points in the image with obvious defects and the standard image without defects, and mark the identified defect feature points in the image with obvious defects and the standard image without defects according to the inductive operation of the obtained editing control. After the marking is completed, a defect set is formed and output to the terminal.

2. The method for detecting surface defects in foamed materials according to claim 1, characterized in that: In S101, if the surface defect detection device acquires the summarization operation of the editing control, and then uses the camera sensor to scan the area of ​​the foam material to be tested to acquire a set of RGB color images, it includes: The summary request for editing controls is sent by a preset network receiving terminal, and the summary request for editing controls is used to request the establishment of a connection with the surface defect detection device; If the summarization request of the editing control sent by the terminal contains a summarization set of editing controls, then the terminal is connected according to the summarization request of the editing control. After the connection is completed, the RGB color image set is obtained by scanning the area of ​​the foaming material to be tested using a camera sensor. The camera sensor can be any type of video camera or infrared camera.

3. The method for detecting surface defects in foamed materials according to claim 2, characterized in that: In S102, the acquired RGB color image set is subjected to image augmentation and adaptation preprocessing. The image augmentation and adaptation preprocessing features include hue, saturation, brightness, contrast changes, image flipping, rotation, image displacement, and center segmentation.

4. The method for detecting surface defects in foamed materials according to claim 3, characterized in that: In S1032, inconspicuous defects include, but are not limited to, microcracks and minor deformations or misalignments.

5. The method for detecting surface defects in foamed materials according to claim 4, characterized in that: In S10323, the new localized region is enlarged and cropped to capture more detailed information, making inconspicuous features in small image regions more apparent, including: The localized small image region is obtained, the edge contour of the small image region is identified and the edge contour feature is generated, and a background occlusion region is generated outside the edge contour feature to shield part of the background of the small image region. Identify and mark areas with high chromatic aberration within the edge contour features. Then, magnify and crop the marked areas with high chromatic aberration to capture more detailed information.

6. A surface defect detection system for foamed materials, characterized in that: A surface defect detection system for performing the method for detecting surface defects in foamed materials according to any one of claims 1-5, the surface defect detection system comprising: The image acquisition unit is used to acquire RGB color images of the area of ​​the foam material to be tested; An image augmentation processing unit is used to improve the clarity of the RGB color image; The coarse registration unit is used to detect whether there are obvious defects in the RGB color image; The fine registration unit is used to detect subtle defects in RGB color images. The fine registration unit includes a positioning module for locating all inconsistent small image regions; The background masking module masks a portion of the background in a small image area by generating a background masking region, thus preserving the features of inconspicuous defect areas. The zoom and crop module is used to zoom in and crop small image areas with masked backgrounds, so as to extract less obvious defect feature information more effectively.

Images