Surface Roughness Measure Based on Average Texture Cycle

This paper presents a testing method of surface roughness for the surface with straight features using image processing. This approach extracts the average texture cycle to describe the surface roughness of mechanical accessory. After surface images of the standard blocks with different roughness levels are collected, the feature value of the average texture cycle which is processed by Matlab software is saved into the texture database. The same method is used to get the feature value of average texture cycle for the surface image of unknown roughness level. Then the surface roughness value of mechanical accessory is gotten by comparing with the feature values of the average texture cycle in the texture database, according to the nearest rule. This method promises higher identification accuracy and lower light source requirement. More so, it can be used effectively when the surface of mechanical accessory has parts of defects.