Research On Methods Of Imaging Optimization And Correction For Heavy Rail Multi-curvature Aggregate Surface

In this dissertation,aiming at the problem of uneven illumination in heavy rail surface imaging based on machine vision,with the multi-curvature aggregate surface characteristics of heavy rail and the theory of machine vision,from two aspects of the optimal configuration of imaging optical path for heavy rail multi-curvature aggregate surface and the correction of curve surface images,the relevant platform was built,the methods of imaging optimization and image correction were designed,and a theoretical and practical basis was provided to improve the defect imaging quality for heavy rail surface detection.The main works of this dissertation are as follows:(1)According to the multi-plane and multi-curve surface characteristics of heavy rail,the optical transfer law and influencing factors of heavy rail surface imaging were analyzed,the heavy rail surface detection scheme was designed,and the hardware platform of heavy rail surface imaging was set up.The optimal experiment on illumination angle was designed,based on the image definition evaluation,the lights layout type and the optimized illuminating angle were determined and an experimental method of imaging optimization for heavy rail surface was provided.(2)By analyzing the characteristics of uneven illumination images of heavy rail curve surface,a correction method was proposed based on image pixel line mean ratio.Comparing with traditional image correction methods,this method had the most uniform gray scale distribution,and could solve the uneven illumination problem of heavy rail curve surface image effectively.(3)Through further study on image filtering and edge segmentation algorithms for heavy rail surface defect image,Gaussian filter method was used to smooth the heavy rail image,and the methods of threshold segmentation and edge detection were compared and analyzed by segmenting the defect image.The results showed that Canny edge detection algorithm could meet the using requirement after the optimization of imaging quality,and the difficulty of defect extraction algorithm was reduced.