Detection method of grinding surface roughness based on image definition evaluation

Abstract

A novel method to detect the grinding surface roughness based on image definition was proposed to solve the problem that the current machine vision detected roughness mainly by adopting image grey value information for statistical analysis, not making full use of color information, and also ignoring the problem of the subjective evaluation of human visual system. According to the phenomenon that the image definition formed on the different grade roughness surface of different color pieces is different, the article built a relational model between resolution and roughness by using two resolution evaluation algorithms, which including the entropy function evaluation algorithm and the color image evaluation algorithm based on color correlation, respectively, to demonstrate the feasibility of the detection method proposed. The experimental results show that the detection method is feasible, the relevance between resolution and roughness is strong, the resolution is decreased while the roughness is increased, and the color image evaluation algorithm based on color correlation is better sensitive. Meanwhile, the idea of the proposed method conforms to the subjective evaluation of human eye vision system, the method combined resolution algorithm with subjective evaluation can detect the whole surface profile roughness of workpiece on-line quickly.