A three-dimensional inspection system for high temperature steel product surface sample height using stereo vision and blue encoded patterns

Vision-based surface inspection system is extremely important research hotspot in the field of steel surface defect inspection. However, most of the vision-based surface inspection systems focus on the two-dimensional image processing, three-dimensional height information acquisition is still a major challenge, especially for high temperature steel product. To achieve the surface sample height information of high temperature steel product sample, a three dimensional inspection system for static object is designed in this paper. In this system, physical filter and digital filter methods are combined to obtain the images of steel product. The stereo rectification and column encoded patterns are employed to achieve the surface sample height information. Furthermore, the graphics processing unit (GPU) is used to improve the computation speed. Seven steel product surface samples are picked to verify the effectiveness of the proposed method. The actual experimental results of the proposed method are very close to the true height, that is, the smallest error is less than 1 mm and the average error is less than 2 mm.