Connector tiny defect detection method and system based on multi-wavelength optical imaging

Abstract

The invention relates to an industrial visual inspection technology, discloses a method and a system for detecting small defects of a connector based on multi-wavelength optical imaging, and solves the problem of low detection precision of the small defects in a traditional optical defect detection scheme. According to the method, firstly, a multi-wavelength light source is adopted to irradiate the surface of the connector under different wavelengths; then respectively acquiring images of the surface of the connector under different wavelengths; preprocessing the collected surface images of the connector under different wavelengths; carrying out image fusion processing on the preprocessed connector surface images under different wavelengths; carrying out defect detection analysis on the fused image by adopting an image detection algorithm, and extracting defect information; and finally, according to the extracted defect information, a deep learning algorithm is adopted to classify defects, whether specification requirements are met or not is judged, and a detection result is output. The method is suitable for micro-defect detection of precise elements such as connectors and the like.

Description

Technical Field

[0001] The present invention relates to industrial visual inspection technology, and in particular to a method and system for detecting tiny defects of connectors based on multi-wavelength optical imaging. Background Art

[0002] Connectors are an indispensable and important component in electronic devices, and their quality directly affects the performance and reliability of the equipment. During the production process of connectors, small defects such as cracks, scratches, bubbles, etc. often appear, which have a significant impact on the reliability and long-term performance of the connector.

[0003] In traditional technologies, optical defect detection of connectors mostly relies on the acquisition and processing of visible light images. However, due to the reflection and scattering characteristics of light, tiny defects are often difficult to clearly distinguish under traditional visible light, especially in the case of uneven lighting and complex background. Tiny defects are more likely to be ignored, resulting in missed detection or false detection.

[0004] Some detection methods also use ultraviolet or infrared imaging to improve detection accuracy. However, due to the limitations of single-wavelength light sources, these methods still have the problem of insufficient response differences to different types of defects, resulting in low detection accuracy. This problem is particularly prominent when facing connectors with complex materials and surface textures.

[0005] As connectors become increasingly smaller and more sophisticated, traditional optical defect detection methods can no longer meet the needs of high-precision detection of tiny defects. Summary of the invention

[0006] The technical problem to be solved by the present invention is to propose a method and system for detecting tiny defects in connectors based on multi-wavelength optical imaging, so as to solve the problem of low accuracy in detecting tiny defects in traditional optical defect detection schemes.

[0007] The technical solution adopted by the present invention to solve the above technical problems is:

[0008] On the one hand, the present invention provides a method for detecting connector micro-defects based on multi-wavelength optical imaging, comprising:

[0009] S1. Use a multi-wavelength light source to illuminate the connector surface at different wavelengths;

[0010] S2, collecting images of the connector surface at different wavelengths respectively;

[0011] S3, preprocessing the collected connector surface images at different wavelengths;

[0012] S4, performing image fusion processing on the pre-processed connector surface images at different wavelengths;

[0013] S5. Use image detection algorithm to perform defect detection and analysis on the fused image to extract defect information;

[0014] S6. Based on the extracted defect information, a deep learning algorithm is used to classify the defects, determine whether they meet the specification requirements, and output the test results.

[0015] Furthermore, the method further comprises the steps of:

[0016] S7. Visualize the test results and control the production line to sort out unqualified connectors or send an alarm signal to the operator.

[0017] Furthermore, in step S3, the preprocessing includes but is not limited to denoising, illumination compensation and image enhancement.

[0018] Furthermore, in step S4, the fusion algorithm used in the image fusion processing includes but is not limited to a multi-scale fusion method based on deep learning.

[0019] Furthermore, in step S5, the defect information includes: shape, size and position information of the defect.

[0020] Furthermore, in step S5, the image detection algorithm includes but is not limited to the Canny edge detection algorithm.

[0021] Furthermore, in step S6, the deep learning method includes but is not limited to Transformer or deep convolutional neural network DCNN.

[0022] On the other hand, the present invention provides a connector micro-defect detection device based on multi-wavelength optical imaging, comprising:

[0023] A multi-wavelength light source module is used to illuminate the connector surface at different wavelengths;

[0024] An image acquisition module, used for respectively acquiring images of the connector surface at different wavelengths;

[0025] A preprocessing module, used for preprocessing the collected connector surface images at different wavelengths;

[0026] An image fusion module, used for performing image fusion processing on the pre-processed connector surface images at different wavelengths;

[0027] Image analysis module, used to use image detection algorithm to perform defect detection and analysis on the fused image and extract defect information;

[0028] The detection output module is used to classify defects based on the extracted defect information using a deep learning algorithm, determine whether they meet the specification requirements, and output the detection results.

[0029] Furthermore, the device also includes:

[0030] The test result feedback module is used to visualize the test results, control the production line to sort out unqualified connectors, or send an alarm signal to the operator.

[0031] The beneficial effects of the present invention are:

[0032] (1) The present invention is based on multi-wavelength optical imaging technology, which can make full use of the response differences of different wavelengths to defects, enhance the contrast between tiny defects and the background, and make tiny defects easier to identify in complex environments, thereby significantly improving the detection accuracy of tiny defects.

[0033] (2) The present invention can effectively suppress background noise, reduce false detection rate, and improve the accuracy of defect recognition through image fusion of multi-wavelength optical images.

[0034] (3) The present invention combines image processing algorithms with deep learning technology, which can not only accurately detect tiny defects, but also accurately classify and judge the defects, greatly improving the detection efficiency and quality control level in the production process.

[0035] Based on the above, the solution of the present invention is particularly suitable for production lines of precision components such as connectors, can effectively improve product quality, ensure efficient automation of the production process, reduce the risks and costs of manual operations, and has important application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a flow chart of a method for detecting connector micro-defects based on multi-wavelength optical imaging in Example 1 of the present invention;

[0037] Figure 2 This is a structural diagram of a connector micro-defect detection device based on multi-wavelength optical imaging in Example 2 of the present invention. DETAILED DESCRIPTION

[0038] The present invention aims to provide a method and system for detecting tiny defects of connectors based on multi-wavelength optical imaging, so as to solve the problem of low accuracy of detecting tiny defects in traditional optical defect detection schemes. The core idea is to introduce multi-wavelength optical imaging technology, and use different reflection and penetration characteristics of different wavelength light sources to enhance the contrast of tiny defects under different lighting conditions, so that tiny defects are easier to identify in complex environments; moreover, after pre-processing the images of different wavelength imaging, a comprehensive decision of multi-wavelength fusion is adopted, so that tiny defects can be detected with a lower false detection rate in more complex connector surface patterns or textures, so as to ensure that deep and fine coating cracks, tiny metal particles and other types of hidden defects can be discovered in time; finally, the image processing algorithm and deep learning technology are combined to automatically extract and accurately classify the defect information, and based on the classification results, it can be judged whether the product meets the specification requirements, and the unqualified products can be automatically sorted by controlling the production line or sending alarm signals to the operator, etc., thereby greatly improving the detection efficiency and quality control level in the production process.

[0039] In terms of specific implementation, the present invention first uses a multi-wavelength light source to illuminate the connector surface; secondly, uses a high-resolution camera to capture images at different wavelengths and performs preprocessing to improve image quality; then, the images at different wavelengths are fused to highlight tiny defects by enhancing the contrast between the defects and the background; then, an image processing algorithm is used to detect and analyze defects, and extract information such as the shape, size, and position of the defects; finally, a deep learning algorithm is used to classify defects, and based on the test results, it is determined whether they meet the specification requirements, and the qualified or unqualified test results are output, the test results are visualized, and the production line is controlled to sort unqualified connectors, or an alarm signal is sent to the operator.

[0040] The scheme of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0041] Example 1

[0042] This embodiment provides a method for detecting connector micro defects based on multi-wavelength optical imaging. Figure 1 , which includes the following implementation steps:

[0043] S1. Multi-wavelength light source irradiation:

[0044] In this step, a multi-wavelength light source is used to illuminate the connector surface at different wavelengths. The multi-wavelength light source has multiple bands such as ultraviolet light, visible light, and near-infrared light. Since light in different bands has different reflection and penetration characteristics, the contrast of tiny defects under different lighting conditions can be enhanced. For example, ultraviolet light has a high reflectivity for microcracks, visible light is suitable for observing surface morphology, and near-infrared light can reveal bubbles, welding defects, etc. The use of a multi-wavelength light source can not only effectively improve the contrast and visibility of tiny defects, but also enable the system to adapt to connectors of various materials and surface textures, and has stronger environmental adaptability.

[0045] S2. Collection of images at different wavelengths:

[0046] In this step, a high-resolution camera can be used in conjunction with multiple optical filters to capture images of the connector surface at different wavelengths.

[0047] S3. Preprocess the collected images:

[0048] In this step, the collected connector surface images at different wavelengths are preprocessed; the preprocessing includes but is not limited to denoising, illumination compensation and image enhancement; image preprocessing can improve image quality and the appearance of defect features, laying the foundation for accurate analysis of subsequent images.

[0049] S4, perform image fusion processing:

[0050] In this step, the pre-processed connector surface images at different wavelengths are fused, which uses the difference in the response of different wavelengths to defects to enhance the contrast between tiny defects and the background. For example, ultraviolet images can highlight cracks, infrared images can reveal internal defects, and image fusion can comprehensively improve the overall defect recognition effect. Through multi-wavelength image fusion technology, the system can provide more comprehensive defect information from multiple angles, reduce missed detections and false detections, especially in complex backgrounds and uneven lighting environments. Among them, the fusion algorithms used for image fusion processing include but are not limited to multi-scale fusion methods based on deep learning.

[0051] S5. Detection and analysis to extract defect information:

[0052] In this step, an image detection algorithm is used to perform defect detection and analysis on the fused image to extract defect information; wherein, the image detection algorithm includes but is not limited to the Canny edge detection algorithm, and the extracted defect information includes the shape, size and position information of the defect, which can be used for subsequent defect classification and judgment on whether it meets the specification requirements.

[0053] S6. Classify the defects and output the test results:

[0054] In this step, based on the extracted defect information, a deep learning algorithm is used to classify the defects, determine whether they meet the specification requirements, and output the test results. The deep learning algorithm includes but is not limited to Transformer or deep convolutional neural network DCNN.

[0055] S7. Test result feedback:

[0056] In this step, the test results can be visualized, and the production line can be controlled to sort out unqualified connectors, or an alarm signal can be sent to the operator. Based on this, automatic detection, automatic classification and automatic sorting can be achieved, reducing manual intervention and improving production efficiency.

[0057] Example 2

[0058] This embodiment provides a connector micro defect detection device based on multi-wavelength optical imaging, see Figure 2 , which includes: multi-wavelength light source module, image acquisition module, preprocessing module, image fusion module, image analysis module, detection output module, and detection result feedback module. The functions of each module are as follows:

[0059] A multi-wavelength light source module is used to illuminate the connector surface at different wavelengths;

[0060] An image acquisition module, used for respectively acquiring images of the connector surface at different wavelengths;

[0061] A preprocessing module, used for preprocessing the collected connector surface images at different wavelengths;

[0062] An image fusion module, used for performing image fusion processing on the pre-processed connector surface images at different wavelengths;

[0063] Image analysis module, used to use image detection algorithm to perform defect detection and analysis on the fused image and extract defect information;

[0064] The detection output module is used to classify defects based on the extracted defect information using a deep learning algorithm, determine whether they meet the specification requirements, and output the detection results.

[0065] The test result feedback module is used to visualize the test results, control the production line to sort out unqualified connectors, or send an alarm signal to the operator.

[0066] Based on the above detection device, the multi-wavelength light source module is used to illuminate the connector surface at different wavelengths respectively, and then the image acquisition module is used to collect images of the connector surface at different wavelengths; then, the preprocessing module is used to preprocess the collected connector surface images at different wavelengths; and the image fusion module is used to perform image fusion processing on the preprocessed connector surface images at different wavelengths; then, the image analysis module is used to perform defect detection and analysis on the fused images to extract defect information; then, the detection output module is used to classify the defects according to the extracted defect information using a deep learning algorithm, and determine whether it meets the specification requirements, and output the detection results; finally, the detection result feedback module is used to visualize the detection results, and the production line is controlled to sort out unqualified connectors, or an alarm signal is sent to the operator.

[0067] Finally, it should be noted that the above embodiments are only preferred implementations and are not intended to limit the present invention. It should be pointed out that for those skilled in the art, several modifications, equivalent replacements, improvements, etc. can be made without departing from the scope of the present invention and the scope of protection of the claims, and all of these should be included in the protection scope of the present invention.

Claims

1. A method for detecting connector micro-defects based on multi-wavelength optical imaging, characterized in that: include: S1. Use a multi-wavelength light source to illuminate the connector surface at different wavelengths; S2, collecting images of the connector surface at different wavelengths respectively; S3, preprocessing the collected connector surface images at different wavelengths; S4, performing image fusion processing on the pre-processed connector surface images at different wavelengths; S5. Use image detection algorithm to perform defect detection and analysis on the fused image to extract defect information; S6. Based on the extracted defect information, a deep learning algorithm is used to classify the defects, determine whether they meet the specification requirements, and output the test results.

2. A method for detecting connector micro-defects based on multi-wavelength optical imaging as claimed in claim 1, characterized in that: Also includes the steps: S7. Visualize the test results and control the production line to sort out unqualified connectors or send an alarm signal to the operator.

3. A method for detecting connector micro-defects based on multi-wavelength optical imaging as claimed in claim 1, characterized in that: In step S3, the preprocessing includes denoising, illumination compensation and image enhancement.

4. A method for detecting connector micro-defects based on multi-wavelength optical imaging as claimed in claim 1, characterized in that: In step S4, the fusion algorithm used for the image fusion processing includes a multi-scale fusion method based on deep learning.

5. A method for detecting connector micro-defects based on multi-wavelength optical imaging according to any one of claims 1 to 4, characterized in that: The defect information includes: shape, size and location information of the defect.

6. A connector micro-defect detection device based on multi-wavelength optical imaging, characterized in that: include: A multi-wavelength light source module is used to illuminate the connector surface at different wavelengths; An image acquisition module, used for respectively acquiring images of the connector surface at different wavelengths; A preprocessing module, used for preprocessing the collected connector surface images at different wavelengths; An image fusion module, used for performing image fusion processing on the pre-processed connector surface images at different wavelengths; Image analysis module, used to use image detection algorithm to perform defect detection and analysis on the fused image and extract defect information; The detection output module is used to classify defects based on the extracted defect information using a deep learning algorithm, determine whether they meet the specification requirements, and output the detection results.

7. A connector micro-defect detection device based on multi-wavelength optical imaging as claimed in claim 6, characterized in that: The device also includes: The test result feedback module is used to visualize the test results, control the production line to sort out unqualified connectors, or send an alarm signal to the operator.

8. The connector micro-defect detection device based on multi-wavelength optical imaging according to claim 6, characterized in that: The preprocessing module performs preprocessing on the collected connector surface images at different wavelengths, including: denoising, illumination compensation and image enhancement.

9. A connector micro-defect detection device based on multi-wavelength optical imaging as claimed in claim 6, characterized in that: The image fusion module performs image fusion processing on the pre-processed connector surface images at different wavelengths, and the fusion algorithm used includes a multi-scale fusion method based on deep learning.

10. A connector micro-defect detection device based on multi-wavelength optical imaging according to any one of claims 6 to 9, characterized in that: The defect information extracted by the image analysis module includes: shape, size and position information of the defect.

Images