Research On Visual Inspection Technology Of Surface Crack Of Special Steel Bar

As an important industrial raw material,special steel bar plays an important role in China’s iron and steel industry.However,in the process of production,some small cracks often appear near the surface of the bar,which reduce the service life of products and cause economic losses to related enterprises.At present,for the crack defect on the surface of special steel bar,enterprises generally adopt the manual visual inspection method,which leads to low inspection efficiency,high misdetection rate and failure to combine with automatic grinding technology.In this paper,the machine vision detection technology is introduced to study the small crack defect near the surface of special steel bar.The main research contents include:Based on the analysis of the detection requirements of the crack on the surface of the bar,a scheme is designed to collect the surface image of the bar by a linear array camera with a fixed line frequency based on the rotation of the bar,so as to avoid the problem of the different brightness of the center and edge of the surface of the cylinder caused by the curvature of the light source.At the same time,the parameters of linear camera,light source,lens and software are selected.Taking the surface crack image of the bar as the research object,the images with consistent horizontal and vertical resolution are collected.The region of interest(ROI)is extracted from the bar surface image,and the filtering effect is evaluated from objective and subjective aspects by comparing the mean filter,median filter and Gaussian filter.The bilateral filter is adopted to remove the surface noise of the image,and the crack edge details are preserved while the noise of the bar surface is removed.In addition,the image of the bar is enhanced and the bright spot area is removed to improve the contrast between the crack and the background area.The traditional Canny edge detection algorithm is improved,and the maximum inter-class variance method is used to determine the high and low thresholds of Canny edge detection in different bar images,which improves the precision of edge detection.According to the geometric features of the selected region,such as area,aspect ratio and angle with the horizontal direction,the real crack is screened out and the crack defect location information is coordinated.By using Halcon image algorithm package and C# language for joint programming,a concise interactive interface is designed,the software function module design is completed,and the surface image collection,processing,display and storage of special steel bar are realized.The experimental results show that the improved Canny edge detection algorithm has the least interference area when segmenting the crack area on the bar surface,and the information of the crack edge remains intact.Through the experiment on the stability of the algorithm of the crack defect detection system,the correct recognition rate of the image with crack reaches 93.75%,the algorithm has low misdetection rate and short running time,which meets the design requirements of the system.