Machine vision inspection of roughness defects in opaque board surface

A machine visual method is introduced to inspect roughness defects from the opaque board surface. The new method uses the line-structured light technology with direct triangulation structure to obtain three-dimensional (3D) information. The captured light stripe image modulated by the board surface is processed firstly with the Steger algorithm to extract the center line of the light strip with sub-pixel level. The key parameter used in the Steger algorithm is obtained by estimating the width of light strip. Furthermore, the Zhang Zhengyou's method is used to calibrate the camera parameters, and the 3D depth information is calibrated through the light plane fitting method combining with the principal component analysis (PCA). As a result, the 3D coordinate information corresponding to the center line of the light stripe on the measured surface can be obtained. Finally, the average deviation of profile arithmetic (R_a index) is employed to evaluate the roughness defect of board surface. The experimental results indicate that the method proposed in the paper can achieve a detecting accuracy of 0.02mm (R_a), which can meet the needs of high-quality wooden board for roughness defect inspection. Compared with the traditional contact detecting method and manual method, the proposed method has high detecting efficiency, low cost, reliable detecting result, and holds a good engineering application prospect.