Welding seam defect detection method and system capable of resisting smog and strong light interference based on electric arc welding environment and medium

Abstract

The invention provides a welding seam defect detection method and system resistant to smoke and strong light interference based on an electric arc welding environment and a medium, and the method comprises the steps: obtaining a shot image of the electric arc welding environment, and carrying out the enhancement processing of the shot image, and obtaining an enhanced image; analyzing light source projection information based on the enhanced image, and comparing the light source projection information with set projection information to obtain light source transmittance; judging whether the light source transmissivity is greater than or equal to a set transmissivity threshold value; if the light source transmissivity is smaller than the light source transmissivity threshold value, performing smoke removal processing on the enhanced image based on a smoke removal model to obtain a smog-free image; analyzing the smog-free image based on the weld defect detection model, and outputting weld defect information; if the light source transmissivity is greater than or equal to the set light source transmissivity, analyzing the enhanced image based on a weld defect model, and outputting weld defect information; the smoke concentration is judged by analyzing the light source projection information of the electric arc welding environment, so that the image is subjected to smoke removal treatment, and the weld defect detection precision is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of weld defects, in particular to a weld defect detection method and system resisting strong light interference of smoke based on an arc welding environment and a medium. BACKGROUND

[0002] With the popularization of industrial automation and intelligent manufacturing, arc welding, as an efficient and reliable metal connection process, is widely used in the fields of automobile manufacturing, rail transportation, energy and power, and shipbuilding. In order to achieve high-quality and high-consistency welding quality, the industry increasingly relies on machine vision systems to monitor and feedback data in real time during the welding process.

[0003] However, the arc welding site environment is extremely complex, and the following key interferences are faced during image acquisition: 1) strong arc light interference, The strong light radiation (including ultraviolet, visible and infrared) generated by the arc has extremely high brightness, far exceeding the dynamic range of the camera image sensor, resulting in serious overexposure and bright spot saturation of the image, which hides the details of the molten pool and the weld; 2) high-density smoke shielding, Metal vapor condenses into particulate smoke, which is unevenly distributed and randomly flows, causing severe light scattering and image blurring, which causes the image contrast to drop sharply; 3) spatter interference and infrared radiation, Strong infrared interference is generated by molten droplet spatter and high-temperature molten pool, which affects the clarity of visible light images and may cause damage to the sensor; 4) dramatic changes in time scale, The arc fluctuation in the arc welding process is fast and unstable, and the transient behavior of the arc starting / extinguishing requires a very high response speed and shutter control for the vision system.

[0004] The above interferences seriously affect the imaging quality of the machine vision system, resulting in the following consequences: 1) unable to clearly observe the molten pool shape and weld contour; 2) low image contrast and unclear edge features; 3) difficulty in accurately identifying welding defects; 4) increased error of AI algorithm training samples. SUMMARY

[0005] The purpose of the embodiments of the present application is to provide a weld defect detection method and system resisting strong light interference of smoke based on an arc welding environment and a medium. By analyzing the source transmittance, the enhanced image is analyzed based on the weld defect model, and the weld defect information is output. By analyzing the light source projection information of the arc welding environment to judge the smoke concentration, the image is de-smoked, and the weld defect detection accuracy is improved.

[0006] The embodiment of the present application also provides a welding seam defect detection method resisting smoke strong light interference based on an electric arc welding environment, comprising the following steps: An image of the electric arc welding environment is acquired, and the image is subjected to enhancement processing to obtain an enhanced image; Light source projection information is analyzed based on the enhanced image, and the light source projection information is compared with set projection information to obtain light source transmittance; It is judged whether the light source transmittance is greater than or equal to a set transmittance threshold value; If the light source transmittance is less than the transmittance threshold value, the enhanced image is subjected to smoke removal processing based on a smoke removal model to obtain a smoke-free image; The smoke-free image is analyzed based on a welding seam defect detection model to output welding seam defect information; If the light source transmittance is greater than or equal to the set transmittance, the enhanced image is analyzed based on a welding seam defect model to output welding seam defect information.

[0007] Optionally, in the welding seam defect detection method resisting smoke strong light interference based on the electric arc welding environment, the image of the electric arc welding environment is acquired, and the image is subjected to enhancement processing to obtain the enhanced image, and specifically comprising the following steps: A high dynamic range camera matched based on the electric arc welding environment is selected to capture the image to obtain the captured image; A segmentation region value is set, and the captured image is subjected to block processing based on the segmentation region value to obtain a plurality of image blocks; Pixel values of each image block are acquired, and the pixel values of each image block are subjected to histogram equalization processing to obtain a processed histogram; Image contrast is analyzed based on the processed histogram, a contrast threshold value is set, and the image contrast is compared with the contrast threshold value; If the image contrast is less than the contrast threshold value, the corresponding image block is subjected to enhancement processing to obtain the enhanced image.

[0008] Optionally, in the welding seam defect detection method resisting smoke strong light interference based on the electric arc welding environment, the captured image is subjected to enhancement processing to obtain the enhanced image, and the method further comprises the following steps: Pixel values of the image block are analyzed based on the processed histogram, and exposure values of each image block pixel are analyzed based on the pixel values of the image block; An exposure threshold interval is set, and the exposure values are compared with the exposure threshold interval, wherein the exposure threshold interval comprises a lower limit value of the exposure threshold interval and an upper limit value of the exposure threshold interval; If the exposure values are in the exposure threshold interval, the pixel points are determined as normal pixel points, and an exposure mean value of the normal pixel points is calculated; If the exposure value is less than the lower limit value of the exposure threshold interval, the pixel point is determined as an underexposed pixel point, and if the exposure value is greater than the upper limit value of the exposure threshold interval, the pixel point is determined as an overexposed pixel point; The underexposed pixel points and the overexposed pixel points are filled based on the exposure mean value of the normal pixel points to obtain a processed shooting image, and the processed shooting image is subjected to noise elimination based on a median filter algorithm to obtain an enhanced image.

[0009] Optionally, in the welding seam defect detection method based on the arc welding environment and resisting smoke strong light interference, the light source projection information is analyzed based on the enhanced image, the light source projection information is compared with the set projection information to obtain a light source transmittance, and specifically includes the following steps: The light source projection region is extracted by analyzing the enhanced image based on a threshold segmentation algorithm and an edge detection algorithm; The average gray value and the peak gray value of the light source region are calculated to obtain a light source intensity feature; The position, width, curvature and fracture of the light strip of the light source region are analyzed to obtain a geometric feature; The coordinate distribution, symmetry and dispersion of the light source in the enhanced image are analyzed to obtain a spatial distribution feature; In a clean environment without welding smoke and arc interference, a light source projection image is shot as a reference, and an ideal intensity threshold, an ideal geometric shape and an ideal spatial distribution coordinate of the light source are set; The light source projection information is obtained based on the light source intensity feature, the geometric feature and the spatial distribution feature, the set projection information is obtained based on the ideal intensity threshold, the ideal geometric shape and the ideal spatial distribution coordinate, and the light source projection information is compared with the set projection information to obtain a light source projection rate.

[0010] Optionally, in the welding seam defect detection method based on the arc welding environment and resisting smoke strong light interference, the smoke removal model construction method is as follows: The image containing smoke is collected based on the arc welding environment, and the image without smoke is matched as a data label to construct a data set; The data set is divided into a training set and a verification set according to a set proportion; The model architecture is constructed, the model architecture is iteratively trained based on the training set, and a loss value is calculated; The loss value is compared with a set loss threshold value; If the loss value is greater than or equal to the set loss threshold value, correction information is generated, and the hyperparameters of the model architecture are dynamically adjusted based on the correction information; If the loss value is less than or equal to the set loss value, an initial smoke removal model is generated, the initial smoke removal model is verified based on the verification set, the hyperparameters of the model architecture are dynamically adjusted based on the verification result, and a final smoke removal model is obtained.

[0011] Optionally, in the welding seam defect detection method based on the arc welding environment and resisting smoke strong light interference, the smoke-free image is analyzed based on the welding seam defect detection model, and welding seam defect information is output. Specifically, the method comprises the following steps: Various welding seam images are collected in a smoke-free environment, wherein the various welding seam images include normal welding seam images, pore images, crack images, incomplete fusion images, and undercut images. The welding seam types are analyzed based on the various welding seam images, and the welding seam images of different welding seam types are labeled. The welding seam images of different welding seam types are combined at different proportions to obtain a training set. The initial model is iteratively trained based on the training set to obtain a training result. It is determined whether the training result converges. If the training result converges, a welding seam defect detection model is generated, and a smoke-free image is analyzed based on the welding seam defect detection model to obtain welding seam defect information. If the training result does not converge, the proportion of the welding seam images of different welding seam types is adjusted, and the initial model is trained again.

[0012] In a second aspect, the embodiments of the present application provide a welding seam defect detection system based on an arc welding environment and resisting smoke strong light interference. The system comprises a memory and a processor. The memory comprises a program of a welding seam defect detection method based on an arc welding environment and resisting smoke strong light interference. When the program of the welding seam defect detection method based on an arc welding environment and resisting smoke strong light interference is executed by the processor, the following steps are implemented: A shooting image of an arc welding environment is obtained, and the shooting image is enhanced to obtain an enhanced image. Light source projection information is analyzed based on the enhanced image, and the light source projection information is compared with set projection information to obtain a light source transmittance. It is determined whether the light source transmittance is greater than or equal to a set transmittance threshold. If the light source transmittance is less than the light source transmittance threshold, a smoke-free image is obtained by performing smoke removal processing on the enhanced image based on a smoke removal model. A welding seam defect detection model is used to analyze the smoke-free image, and welding seam defect information is output. If the light source transmittance is greater than or equal to the set light source transmittance, a welding seam defect model is used to analyze the enhanced image, and welding seam defect information is output.

[0013] Optionally, in the welding seam defect detection system based on the arc welding environment and resisting smoke strong light interference, a shooting image of an arc welding environment is obtained, and the shooting image is enhanced to obtain an enhanced image. Specifically, the method comprises the following steps: A high dynamic camera matched based on the arc welding environment is selected to take a photograph, and a photographed image is obtained. A segmentation region value is set, and the photographed image is processed in blocks based on the segmentation region value, and a plurality of image blocks are obtained. Pixel values of each image block are obtained, and the pixel values of each image block are processed by histogram equalization, and a processed histogram is obtained. The image contrast is analyzed based on the processed histogram, a contrast threshold is set, and the image contrast is compared with the contrast threshold. If the image contrast is less than the contrast threshold, the corresponding image block is processed for enhancement, and an enhanced image is obtained.

[0014] Optionally, in the welding seam defect detection system against smoke and strong light interference based on the arc welding environment, the photographed image is processed for enhancement, and an enhanced image is obtained, and the system further comprises: The pixel values of the image block are analyzed based on the processed histogram, and the exposure value of each pixel point of each image block is analyzed based on the pixel values of the image block. An exposure threshold interval is set, and the exposure value is compared with the exposure threshold interval, the exposure threshold interval comprising a lower limit value of the exposure threshold interval and an upper limit value of the exposure threshold interval. If the exposure value is in the exposure threshold interval, the pixel point is determined as a normal pixel point, and the exposure mean value of the normal pixel point is calculated. If the exposure value is less than the lower limit value of the exposure threshold interval, the pixel point is determined as an underexposed pixel point, and if the exposure value is greater than the upper limit value of the exposure threshold interval, the pixel point is determined as an overexposed pixel point. The underexposed pixel point and the overexposed pixel point are processed for filling based on the exposure mean value of the normal pixel point, and a processed photographed image is obtained, and the processed photographed image is processed for noise removal based on a median filter algorithm, and an enhanced image is obtained.

[0015] In a third aspect, the embodiments of the present application further provide a computer readable storage medium, wherein a welding seam defect detection method program against smoke and strong light interference based on an arc welding environment is stored in the computer readable storage medium, and when the welding seam defect detection method program against smoke and strong light interference based on the arc welding environment is executed by a processor, the steps of the welding seam defect detection method against smoke and strong light interference based on the arc welding environment are implemented.

[0016] As can be seen from the above, the welding seam defect detection method, system and medium provided by the embodiment of the application based on the arc welding environment and capable of resisting smoke and strong light interference, by acquiring a shooting image of the arc welding environment, performing enhancement processing on the shooting image, and obtaining an enhanced image, analyzing light source projection information based on the enhanced image, comparing the light source projection information with set projection information, obtaining light source transmittance, judging whether the light source transmittance is greater than or equal to a set transmittance threshold, if the light source transmittance is less than the light source transmittance threshold, performing smoke removal processing on the enhanced image based on a smoke removal model to obtain a smoke-free image, analyzing the smoke-free image based on a welding seam defect detection model to output welding seam defect information, if the light source transmittance is greater than or equal to the set light source transmittance, analyzing the enhanced image based on the welding seam defect model to output the welding seam defect information, and judging the smoke concentration by analyzing the light source projection information of the arc welding environment, thereby performing smoke removal processing on the image and improving the welding seam defect detection precision. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments of the present application. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0018] Figure 1 The flowchart of the welding seam defect detection method based on the arc welding environment and capable of resisting smoke and strong light interference provided by the embodiment of the present application is shown in FIG. 1. Figure 2 The shooting image enhancement processing flowchart of the welding seam defect detection method based on the arc welding environment and capable of resisting smoke and strong light interference provided by the embodiment of the present application is shown in FIG. 2. Figure 3 The shooting image exposure value filling method flowchart of the welding seam defect detection method based on the arc welding environment and capable of resisting smoke and strong light interference provided by the embodiment of the present application is shown in FIG. 3. DETAILED DESCRIPTION

[0019] The technical solutions of the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of the present application.

[0020] It should be noted that similar reference numerals and letters refer to like items throughout the several views, and that, as such, no further definitions and explanations of the same are required in the subsequent views. Also, in the description of the present application, the terms "first", "second", etc. are used merely to distinguish descriptions and can not be understood as indicating or implying relative importance.

[0021] Referring to Figure 1 , Figure 1 is a flowchart of a welding seam defect detection method based on an arc welding environment and resisting smoke strong light interference in some embodiments of the present application. The welding seam defect detection method based on the arc welding environment and resisting the smoke strong light interference is used in a terminal device. The welding seam defect detection method based on the arc welding environment and resisting the smoke strong light interference comprises the following steps: S101, a shooting image of the arc welding environment is acquired, and the shooting image is subjected to enhancement processing to obtain an enhanced image; S102, light source projection information is analyzed based on the enhanced image, and the light source projection information is compared with set projection information to obtain light source transmittance; S103, it is judged whether the light source transmittance is greater than or equal to a set transmittance threshold value; S104, if the light source transmittance is less than the light source transmittance threshold value, a smoke removal model is used to remove smoke from the enhanced image to obtain a smoke-free image, and a welding seam defect detection model is used to analyze the smoke-free image to output welding seam defect information; S105, if the light source transmittance is greater than or equal to the set light source transmittance, a welding seam defect model is used to analyze the enhanced image to output welding seam defect information.

[0022] It should be noted that the shooting image is subjected to enhancement processing, a high-resolution image is output in a dynamic welding environment, the image is subjected to smoke removal processing, the image clarity is improved, the small defects in the image are ensured to be recognized, and the defect recognition accuracy is improved.

[0023] Referring to Figure 2 , Figure 2 is a shooting image enhancement processing flowchart of a welding seam defect detection method based on an arc welding environment and resisting smoke strong light interference in some embodiments of the present application. According to the embodiment of the present application, a shooting image of the arc welding environment is acquired, and the shooting image is subjected to enhancement processing to obtain an enhanced image. Specifically, the enhancement processing comprises the following steps: S201, a high dynamic camera matched based on the arc welding environment is selected to shoot to obtain a shooting image; S202, a segmentation region value is set, and the shooting image is subjected to block processing based on the segmentation region value to obtain a plurality of image blocks; S203, obtain the pixel value of each image block, perform histogram equalization processing on the pixel value of each image block, and obtain a processed histogram; S204, analyze the image contrast based on the processed histogram, set a contrast threshold, and compare the image contrast with the contrast threshold; S205, if less than the contrast threshold, the corresponding image block is enhanced, and an enhanced image is obtained.

[0024] It should be noted that the captured image is regionally segmented, and the segmented image block is processed separately to improve the processing effect.

[0025] Please refer to Figure 3 , Figure 3 is a flow chart of a shooting image exposure value filling method of a welding seam defect detection method based on arc welding environment and anti-smoke strong light interference. According to the embodiment of the present application, the captured image is enhanced to obtain an enhanced image, which further comprises: S301, based on the processed histogram, analyze the pixel value of the image block, and analyze the exposure value of each image block pixel based on the pixel value of the image block; S302, set an exposure threshold interval, compare the exposure value with the exposure threshold interval, and the exposure threshold interval includes a lower limit value of the exposure threshold interval and an upper limit value of the exposure threshold interval; S303, if the exposure value is in the exposure threshold interval, it is determined that the pixel point is a normal pixel point, and the exposure mean value of the normal pixel point is calculated; S304, if the exposure value is less than the lower limit value of the exposure threshold interval, it is determined that the pixel point is an underexposed pixel point, and if the exposure value is greater than the upper limit value of the exposure threshold interval, it is determined that the pixel point is an overexposed pixel point; S305, based on the exposure mean value of the normal pixel point, the underexposed pixel point and the overexposed pixel point are filled and processed to obtain a processed captured image, and based on the median filter algorithm, the processed captured image is removed from the noise to obtain an enhanced image.

[0026] It should be noted that the exposure value is analyzed to suppress arc interference, suppress intense arc highlight area, restore the true gray scale of the welding seam area, enhance the penetration ability, and remove the image blur caused by smoke.

[0027] According to the embodiment of the present application, based on the enhanced image analysis of the light source projection information, the light source projection information is compared with the set projection information to obtain the light source transmittance, which specifically comprises: Based on the threshold segmentation algorithm and the edge detection algorithm, the enhanced image is analyzed, and the light source projection area is extracted; Calculate the average gray value and peak gray value of the light source area to obtain the light source intensity feature; The position, width, curvature and breaking of the light stripe of the light source region are analyzed to obtain geometric characteristics; The coordinate distribution, symmetry and dispersion of the light source in the enhanced image are analyzed to obtain spatial distribution characteristics; In a clean environment without welding smoke and arc interference, the light source projection image is shot as a reference to set the ideal intensity threshold, ideal geometric shape and ideal spatial distribution coordinates of the light source; The light source projection information is obtained based on the intensity characteristics, geometric characteristics and spatial distribution characteristics of the light source, and the set projection information is obtained based on the ideal intensity threshold, ideal geometric shape and ideal spatial distribution coordinates, and the light source projection information is compared with the set projection information to obtain the light source projection rate.

[0028] It should be noted that the light source projection information is analyzed by analyzing the intensity characteristics, geometric characteristics and spatial distribution coordinates of the image, and compared with the set projection information, so that the light source transmission rate is accurately analyzed.

[0029] According to the embodiment of the present application, the smoke removal model construction method is as follows: Based on the arc welding environment, the image containing smoke is collected, and the image without smoke is matched as a data label to construct a data set; The data set is divided into a training set and a verification set according to a set proportion; The model architecture is constructed, the model architecture is iteratively trained based on the training set, and the loss value is calculated; The loss value is compared with the set loss threshold; If the loss value is greater than or equal to the set loss threshold, correction information is generated, and the hyperparameters of the model architecture are dynamically adjusted based on the correction information; If the loss value is less than or equal to the set loss value, an initial smoke removal model is generated, the initial smoke removal model is verified based on the verification set, the hyperparameters of the model architecture are dynamically adjusted based on the verification result, and the final smoke removal model is obtained.

[0030] It should be noted that the smoke removal model (based on dark channel prior) is constructed as follows: Suppose the image model is: , Solve the target image : , Wherein: A: atmospheric light estimation; : projection, estimated by the minimum channel local window; : transmission lower limit, to avoid division by zero.

[0031] High dynamic multi-exposure fusion is as follows: Obtain multiple images Different exposure images, weighted fusion: , Wherein: Generally represents the pixel value of the image obtained under the i-th different exposure condition at the pixel position (x, y), which comes from multiple shots using different exposure parameters to capture more dynamic range information of the scene; Indicates the value of the high dynamic range image (HDR image) at the pixel position (x, y) after weighted fusion, that is, the result image after fusion; : According to the local gradient, the brightness and the saturation dynamic adjustment weight.

[0032] Laplace enhancement and Retinex model are used to sharpen the image and restore the edge, and the details and contrast are improved.

[0033] According to the embodiment of the present application, the smoke-free image is analyzed based on the welding seam defect detection model, and the welding seam defect information is output, specifically including: Collect various welding seam images in a smoke-free environment, and the various welding seam images include normal welding seam images, pore images, crack images, un-melted images and undercut images; Analyze the welding seam type based on various welding seam images, and label the welding seam images of different welding seam types; Combining the welding seam images of different welding seam types in different proportions to obtain a training set; Iteratively training the initial model based on the training set to obtain a training result; Judge whether the training result converges or not; If it converges, a welding seam defect detection model is generated, and the smoke-free image is analyzed based on the welding seam defect detection model to obtain the welding seam defect information; If it does not converge, adjust the proportion of the welding seam images of different welding seam types, and perform secondary training on the initial model.

[0034] It should be noted that the training set is constructed by combining the proportions of the welding seam images of different welding seam types, and the welding seam defect detection model is obtained, so as to accurately identify the welding seam defect type and the welding seam defect position in the image.

[0035] In a second aspect, the embodiment of the present application provides a welding seam defect detection system resisting smoke strong light interference based on an electric arc welding environment, which comprises a memory and a processor, the memory comprising a program of a welding seam defect detection method resisting smoke strong light interference based on an electric arc welding environment, and the program of the welding seam defect detection method resisting smoke strong light interference based on an electric arc welding environment is executed by the processor to realize the following steps: An image of the electric arc welding environment is acquired, and the image is subjected to enhancement processing to obtain an enhanced image; The enhanced image is analyzed based on light source projection information, and the light source projection information is compared with set projection information to obtain light source transmittance; It is determined whether the light source transmittance is greater than or equal to a set transmittance threshold value; If the light source transmittance is less than the transmittance threshold value, the enhanced image is subjected to smoke removal processing based on a smoke removal model to obtain a smoke-free image; The smoke-free image is analyzed based on a welding seam defect detection model to output welding seam defect information; If the light source transmittance is greater than or equal to the set transmittance threshold value, the enhanced image is analyzed based on a welding seam defect model to output welding seam defect information.

[0036] It should be noted that the image is subjected to enhancement processing, which can quickly respond in a dynamic welding environment, output a high-resolution image, and perform smoke removal processing on the image to improve image clarity, ensure that small defects in the image can be recognized, and improve defect recognition accuracy.

[0037] According to the embodiment of the present application, an image of the electric arc welding environment is acquired, and the image is subjected to enhancement processing to obtain an enhanced image, which specifically comprises: A high dynamic camera matched based on the electric arc welding environment is selected to capture an image; A segmentation region value is set, and the image is subjected to block processing based on the segmentation region value to obtain a plurality of image blocks; Pixel values of each image block are acquired, and the pixel values of each image block are subjected to histogram equalization processing to obtain a processed histogram; The image contrast is analyzed based on the processed histogram, a contrast threshold value is set, and the image contrast is compared with the contrast threshold value; If the image contrast is less than the contrast threshold value, the corresponding image block is subjected to enhancement processing to obtain an enhanced image.

[0038] It should be noted that the image is subjected to region segmentation, and the segmented image blocks are processed separately to improve processing effect.

[0039] According to the embodiment of the present application, the image is subjected to enhancement processing to obtain an enhanced image, which further comprises: Based on the processed histogram analysis of the pixel values of the image block, the exposure value of each pixel point of the image block is analyzed based on the pixel values of the image block; An exposure threshold interval is set, and the exposure value is compared with the exposure threshold interval, the exposure threshold interval including a lower limit value of the exposure threshold interval and an upper limit value of the exposure threshold interval; If the exposure value is in the exposure threshold interval, it is determined that the pixel point is a normal pixel point, and the exposure mean value of the normal pixel point is calculated; If the exposure value is less than the lower limit value of the exposure threshold interval, it is determined that the pixel point is an underexposed pixel point, and if the exposure value is greater than the upper limit value of the exposure threshold interval, it is determined that the pixel point is an overexposed pixel point; Based on the exposure mean value of the normal pixel point, the underexposed pixel point and the overexposed pixel point are filled and processed to obtain a processed shooting image, and based on a median filter algorithm, noise is removed from the processed shooting image to obtain an enhanced image.

[0040] It should be noted that by analyzing the exposure value, arc light interference is suppressed, intense arc highlight areas are suppressed, the true gray scale of the weld area is restored, the penetration ability is enhanced, and image blur caused by smoke is removed.

[0041] According to the embodiment of the present application, based on the enhanced image, the light source projection information is analyzed, the light source projection information is compared with the set projection information, and the light source transmittance is obtained, specifically including: Based on a threshold segmentation algorithm and an edge detection algorithm, the enhanced image is analyzed, and the light source projection area is extracted; The average gray value and the peak gray value of the light source area are calculated to obtain the light source intensity feature; The position, width, curvature and fracture of the light stripe of the light source area are analyzed to obtain the geometric feature; The coordinate distribution, symmetry and dispersion of the light source in the enhanced image are analyzed to obtain the spatial distribution feature; In a clean environment without welding smoke and arc interference, a light source projection image is shot as a reference, and the ideal intensity threshold, ideal geometric shape and ideal spatial distribution coordinates of the light source are set; The light source projection information is obtained based on the light source intensity feature, the geometric feature and the spatial distribution feature, the set projection information is obtained based on the ideal intensity threshold, the ideal geometric shape and the ideal spatial distribution coordinates, and the light source projection information is compared with the set projection information to obtain the light source projection rate.

[0042] It should be noted that by analyzing the light source intensity feature, the geometric feature and the spatial distribution coordinates of the image, the light source projection information is analyzed, and compared with the set projection information, the light source transmittance is accurately analyzed.

[0043] According to the embodiment of the present application, the smoke removal model construction method is as follows: Based on the arc welding environment, collect the image containing fog, and pair the fog-free image as the data label to construct the data set; Divide the data set into a training set and a validation set according to a set proportion; Construct a model architecture, iteratively train the model architecture based on the training set, and calculate the loss value; Compare the loss value with the set loss threshold value; If the loss value is greater than or equal to the set loss threshold value, generate correction information, and dynamically adjust the hyperparameters of the model architecture based on the correction information; If the loss value is less than or equal to the set loss value, generate an initial smoke removal model, verify the initial smoke removal model based on the validation set, dynamically adjust the hyperparameters of the model architecture based on the verification result, and obtain the final smoke removal model.

[0044] It should be noted that the smoke removal model (based on dark channel prior) construction method is as follows: Assume that the image model is: , Solve the target image : , Where: A: atmospheric light estimation; : projection, estimated by the minimum channel local window; : transmission lower limit, to avoid division by zero.

[0045] High dynamic multi-exposure fusion is as follows: Obtain multiple images Different exposure images, weighted fusion: , Where: : dynamically adjust the weight according to the local gradient, brightness and saturation.

[0046] Laplace enhancement and Retinex model are used to sharpen the image and restore the edge, and to improve the details and contrast.

[0047] According to the embodiment of the present application, the smoke-free image is analyzed based on the weld defect detection model, and the weld defect information is output, specifically including: Collect various types of weld images in a smoke-free environment, including normal weld images, pore images, crack images, unfused images and undercut images; Analyze the weld type based on various types of weld images, and label the weld images of different weld types; combining the weld images of different weld types in different proportions to obtain a training set; iteratively training the initial model based on the training set to obtain a training result; determining whether the training result converges; if the training result converges, generating a weld defect detection model, and analyzing the smoke-free image based on the weld defect detection model to obtain weld defect information; if the training result does not converge, adjusting the proportions of the weld images of different weld types, and retraining the initial model.

[0048] It should be noted that the training set is constructed by combining the proportions of the weld images of different weld types, and the weld defect detection model is obtained, so as to accurately identify the weld defect type and weld defect position in the image.

[0049] The third aspect of the application provides a computer readable storage medium, the readable storage medium includes a weld defect detection method program based on an arc welding environment and resisting smoke strong light interference, and the weld defect detection method program based on the arc welding environment and resisting the smoke strong light interference is executed by the processor to realize the steps of the weld defect detection method based on the arc welding environment and resisting the smoke strong light interference according to any one of the above.

[0050] The weld defect detection method, system and medium based on the arc welding environment and resisting the smoke strong light interference disclosed by the application, by acquiring a shooting image of the arc welding environment, performing enhancement processing on the shooting image to obtain an enhanced image, analyzing light source projection information based on the enhanced image, comparing the light source projection information with set projection information to obtain light source transmittance, determining whether the light source transmittance is greater than or equal to a set transmittance threshold, if the light source transmittance is less than the light source transmittance threshold, performing smoke removal processing on the enhanced image based on a smoke removal model to obtain a smoke-free image, analyzing the smoke-free image based on a weld defect detection model to output weld defect information, if the light source transmittance is greater than or equal to the set light source transmittance, analyzing the enhanced image based on a weld defect model to output weld defect information, by analyzing the light source projection information of the arc welding environment to determine the smoke concentration, the image is processed to remove the smoke, and the weld defect detection accuracy is improved.

[0051] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other manners. The above described device embodiments are merely exemplary. For example, the division of the units is only a logical function division. There can be another division manner for the actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the components shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0052] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units; they can be located in one place, or distributed on multiple network units; and part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0053] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be realized in the form of hardware or in the form of hardware plus software functional units.

[0054] Those of ordinary skill in the art can understand that all or part of the steps of the above method embodiments can be completed by a program instructing related hardware, and the foregoing program can be stored in a readable storage medium, and when the program is executed, the steps of the method embodiments are executed; and the foregoing storage medium includes mobile storage devices, read-only memories (ROMs), random access memories (RAMs), magnetic discs or optical discs, and various media that can store program codes.

[0055] Alternatively, the integrated units of the present application, if implemented in the form of software functional modules and sold or used as independent products, can also be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present application can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the embodiments of the present application. The foregoing storage medium includes mobile storage devices, ROMs, RAMs, magnetic discs or optical discs, and various media that can store program codes.

Claims

1. A method for detecting weld defects in an arc welding environment that is resistant to interference from smoke and strong light, characterized by, The method comprises the following steps: obtaining a shooting image of an arc welding environment, and performing enhancement processing on the shooting image to obtain an enhanced image; analyzing light source projection information based on the enhanced image, comparing the light source projection information with set projection information, and obtaining light source transmittance; determining whether the light source transmittance is greater than or equal to a set transmittance threshold value; if the light source transmittance is less than the transmittance threshold value, performing smoke removal processing on the enhanced image based on a smoke removal model to obtain a smoke-free image, analyzing the smoke-free image based on a welding seam defect detection model, and outputting welding seam defect information; if the light source transmittance is greater than or equal to the set transmittance threshold value, analyzing the enhanced image based on a welding seam defect model, and outputting welding seam defect information.

2. The method of detecting weld defects in a welding environment based on an electric arc welding environment that is resistant to smoke and strong light interference according to claim 1, characterized in that, The method comprises the following steps: obtaining a shooting image of an arc welding environment, and performing enhancement processing on the shooting image to obtain an enhanced image, specifically comprising: selecting a matched high-dynamic camera based on the arc welding environment to perform shooting, and obtaining a shooting image; setting a segmentation region value, and performing block processing on the shooting image based on the segmentation region value to obtain a plurality of image blocks; obtaining pixel values of each image block, and performing histogram equalization processing on the pixel values of each image block to obtain a processed histogram; analyzing image contrast based on the processed histogram, setting a contrast threshold value, and comparing the image contrast with the contrast threshold value; 3. The method of detecting weld defects in a welding environment based on an electric arc welding environment that is resistant to smoke and strong light interference according to claim 2, characterized in that, if the image contrast is less than the contrast threshold value, performing enhancement processing on the corresponding image block to obtain an enhanced image. The method further comprises the following steps: analyzing pixel values of the image block based on the processed histogram, and analyzing exposure values of each pixel point of each image block based on the pixel values of the image block; setting an exposure threshold interval, and comparing the exposure values with the exposure threshold interval, wherein the exposure threshold interval comprises a lower limit value of the exposure threshold interval and an upper limit value of the exposure threshold interval; if the exposure value is within the exposure threshold interval, determining that the pixel point is a normal pixel point, and calculating an exposure mean value of the normal pixel point; if the exposure value is less than the lower limit value of the exposure threshold interval, determining that the pixel point is an underexposed pixel point, and if the exposure value is greater than the upper limit value of the exposure threshold interval, determining that the pixel point is an overexposed pixel point; 4. The method of detecting weld defects in a welding environment based on an electric arc welding environment that is resistant to smoke and strong light interference according to claim 3, characterized in that, performing filling processing on the underexposed pixel point and the overexposed pixel point based on the exposure mean value of the normal pixel point to obtain a processed shooting image, and performing noise removal on the processed shooting image based on a median filter algorithm to obtain an enhanced image. The method further comprises the following steps: analyzing the light source projection information based on a threshold segmentation algorithm and an edge detection algorithm to extract a light source projection region; calculating an average gray value and a peak gray value of the light source region to obtain light source intensity characteristics; analyzing positions, widths, curvatures, and fracture conditions of light strips of the light source region to obtain geometric characteristics; analyzing coordinate distribution, symmetry, and dispersion of the light source in the enhanced image to obtain spatial distribution characteristics; in a clean environment without welding smoke and arc interference, shooting a light source projection image as a reference, and setting ideal intensity threshold values, ideal geometric shapes, and ideal spatial distribution coordinates of the light source. The light source projection information is obtained based on light source intensity characteristics, geometric characteristics and spatial distribution characteristics, the set projection information is obtained based on an ideal intensity threshold, an ideal geometric shape and ideal spatial distribution coordinates, the light source projection information is compared with the set projection information, and a light source projection rate is obtained.

5. The method of detecting weld defects in a welding environment based on an electric arc welding environment that is resistant to smoke and strong light interference according to claim 4, characterized in that, The smoke removal model is constructed as follows: An image containing smoke is collected based on an arc welding environment, and a non-smoke image is matched as a data label to construct a data set; The data set is divided into a training set and a verification set according to a set proportion; A model architecture is constructed, and the model architecture is iteratively trained based on the training set to calculate a loss value; The loss value is compared with a set loss threshold value; If the loss value is greater than or equal to the set loss threshold value, correction information is generated, and the hyperparameters of the model architecture are dynamically adjusted based on the correction information; If the loss value is less than or equal to the set loss value, an initial smoke removal model is generated, the initial smoke removal model is verified based on the verification set, the hyperparameters of the model architecture are dynamically adjusted based on the verification result, and a final smoke removal model is obtained.

6. The method of detecting weld defects in a welding environment based on an electric arc welding environment that is resistant to smoke and strong light interference according to claim 5, characterized in that, The non-smoke image is analyzed based on the weld defect detection model to output weld defect information, which specifically includes: Various weld images are collected in a non-smoke environment, including normal weld images, pore images, crack images, incomplete fusion images and undercut images; The weld types are analyzed based on the various weld images, and the weld images of different weld types are labeled; The weld images of different weld types are combined at different proportions to obtain a training set; The initial model is iteratively trained based on the training set to obtain a training result; It is judged whether the training result converges; If it converges, a weld defect detection model is generated, and the non-smoke image is analyzed based on the weld defect detection model to obtain weld defect information; If it does not converge, the proportion of the weld images of different weld types is adjusted, and the initial model is trained again.

7. A weld defect detection system based on arc welding environment that is resistant to smoke and strong light interference, characterized in that, The system includes a memory and a processor, the memory includes a program of the anti-smoke strong light interference weld defect detection method based on the arc welding environment, and the program of the anti-smoke strong light interference weld defect detection method based on the arc welding environment is implemented when the processor is executed to realize the following steps: An enhanced image is obtained by enhancing the captured image of the arc welding environment; The light source projection information is analyzed based on the enhanced image, the light source projection information is compared with the set projection information, and the light source transmission rate is obtained; It is judged whether the light source transmission rate is greater than or equal to the set transmission rate threshold value; If it is less than the light source transmission rate threshold value, the enhanced image is de-smoked based on the de-smoke model to obtain a non-smoke image, and the non-smoke image is analyzed based on the weld defect detection model to output weld defect information; If it is greater than or equal to the set light source transmission rate, the enhanced image is analyzed based on the weld defect model to output weld defect information.

8. The arc welding environment based, smoke resistant, strong light interference weld defect detection system of claim 7, wherein, An enhanced image is obtained by enhancing the captured image of the arc welding environment, specifically including: A matched high dynamic camera is selected based on the arc welding environment to capture an image, and the captured image is obtained; A segmentation region value is set, and the captured image is block-processed based on the segmentation region value to obtain a plurality of image blocks; The pixel value of each image block is obtained, and the pixel value of each image block is subjected to histogram equalization processing to obtain a processed histogram; The image contrast is analyzed based on the processed histogram, a contrast threshold is set, and the image contrast is compared with the contrast threshold; If the image contrast is less than the contrast threshold, the corresponding image block is subjected to enhancement processing to obtain an enhanced image.

9. The arc welding environment based, smoke resistant, strong light interference weld defect detection system of claim 8, wherein, The photographed image is subjected to enhancement processing to obtain an enhanced image, and the method further includes: The pixel value of each image block is obtained, and the pixel value of each image block is subjected to histogram equalization processing to obtain a processed histogram; The pixel value of each image block is obtained, and the pixel value of each image block is subjected to histogram equalization processing to obtain a processed histogram; If the image contrast is less than the contrast threshold, the corresponding image block is subjected to enhancement processing to obtain an enhanced image. The photographed image is subjected to enhancement processing to obtain an enhanced image, and the method further includes: The pixel value of each image block is obtained, and the pixel value of each image block is subjected to histogram equalization processing to obtain a processed histogram; 10. A computer-readable storage medium, characterized in that, The pixel value of each image block is obtained, and the pixel value of each image block is subjected to histogram equalization processing to obtain a processed histogram; If the image contrast is less than the contrast threshold, the corresponding image block is subjected to enhancement processing to obtain an enhanced image. The photographed image is subjected to enhancement processing to obtain an enhanced image, and the method further includes: The computer readable storage medium includes a welding seam defect detection method program resistant to smoke and strong light interference based on an electric arc welding environment, and the welding seam defect detection method program resistant to smoke and strong light interference based on the electric arc welding environment is executed by the processor to realize the steps of the welding seam defect detection method resistant to smoke and strong light interference based on the electric arc welding environment as claimed in any one of claims 1 to 6.

Images