Detection Of Slight Deformation Defects On Highly Reflective Surfaces Based On Contrast Enhancement

With the development of technology and the progress of society,industrial products have been developing towards high quality and diversification.Therefore,it is essential to establish a high-precision and high-efficiency inspection system for industrial products on the assembly line,which helps to screen out high-quality industrial products that meet production standards and consumer needs.However,inspection targets with very low contrast,such as dents in steel plates,are often encountered in industrial inspection.Such slight surface deformation defects have low contrast and no obvious edges under uniform light,which are difficult to image.Such defects are always difficult to detect in machine vision system.Therefore,this paper designs a reasonable optical system for the imaging characteristics of minor defects to improve the contrast of defect imaging,and designs a corresponding detection algorithm for the characteristics of high-contrast images of defects,which makes up for the lack of machine vision system in the detection of minor defects.Firstly,an imaging simulation model of minor defects is established in this paper to verify the imaging enhancement effect of stripe light on minor defects.The model explores the effect of defect size on defect imaging.For the minor defects,the imaging effect of different positions of fringe light is investigated and simultaneously the imaging mechanism of defects is analyzed.On this basis,a fringe light active scanning method was proposed to optically improve the contrast of defect imaging and obtain high-contrast defect images to collect a sequence of images containing high-contrast defects.Secondly,a new image synthesis algorithm based on contrast weighting is proposed for sequential images.To solve the problem of reduced contrast of defects in composite images,the contrast between the corresponding pixels of the sequence image and the surrounding area is calculated,and the pixel value of the composite image is selected according to the contrast weight of each pixel,so as to obtain a high-contrast composite image of defects.In order to address the problem that the contrast weight at the light-dark boundary is relatively large,a column mean threshold is introduced to distinguish defects from the light-dark boundary.The algorithm effectively removes the interference of bright and dark striped backgrounds,retains the high contrast of defects,and obtains a composite image containing complete defects.Thirdly,aiming at the problem that the contrast of slight scratches in synthetic images is still low,an adaptive smoothing and morphological difference algorithm is proposed.According to the different characteristics of scratches,noise and background,different window parameters of the Gaussian filters are selected to smooth the noise to the greatest extent while retaining the scratch characteristics.In the morphological difference process,a difference threshold is introduced to distinguish scratched and non-scratched areas.Only the non-scratch area is differentiated,and the gray value of the scratch is retained,resulting in a high-contrast slightly scratched image.The algorithm effectively removes the interference of noise and uneven background,obtains good segmentation results,and realizes detection and identification of slight deformation defects.Finally,the hardware of the experimental system is selected according to the image acquisition scheme,and a slight defect detection system is built to collect images.The results show that the optical system in this paper covers the inspected surface completely and improves the contrast of minor defects effectively.In addition,the adaptive smoothing and morphological difference algorithm in this paper is compared with the common processing algorithm for minor scratches.The experimental results show that the algorithm in this paper can effectively enhance the defect contrast and obtain a good segmentation effect.Compared with other defect detection methods,the defect detection method in this paper significantly improves the detection ability of slight deformation defects on the surface,which can be applied for industrial inspection.In view of the difficulty of detecting minor defects,this paper builds a reasonable optical system based on the imaging characteristics of minor defects,so as to optically improve the contrast of defect imaging and obtain high-contrast defect images.In addition,a slight defect detection algorithm is designed to further enhance the contrast of weak defects from the image.Finally,a slight defect detection experimental system is built for testing.The experimental results show that the slight defect detection scheme in this paper makes up for the deficiency of the machine vision system in the detection of slight deformation defects on the surface.