Focusing on the 3D topographic characteristics of rolling contact fatigue, a reconstruction method of the fatigue surface of roller based on point cloud data was proposed in this research. A 3D laser scanner was used to capture the data of point cloud on the surface of the fatigue roller. The gradient segmentation method was used to achieve segmentation of the fatigue contact surface, and the Kd-Tree algorithm in Statistical Outlier Removal filter was adopted to remove different types of noise. The greedy triangulation and hole repair and reconstruction of the curled point cloud were conducted. The experimental results showed that the segmentation accuracy of the fatigue contact surface was above 97.7%, the curling error rate of point cloud was 0.09%, and the maximum deviation of the reconstructed fatigue roller surface was 0.0199 mm. These methods can be applied to analyze the working conditions of roller specimen and contact fatigue.

This paper proposes a real-time detection method for gear contact fatigue pitting based on machine vision in order to improve the detection accuracy and detection efficiency of specimen fatigue pitting in gear contact fatigue tests and to realize the visualization, quantification, and real-time detection of gear pitting. Under the principle of gear meshing and the shooting principle of a line-scan camera, a test detection system for gear contact fatigue is established, and the optimal centrifugal shooting distance for the gear tooth surface is obtained by analyzing the gear rotation process. In response to the phenomenon of image overlap caused by the inconsistency between the speed of each point on the gear tooth profile and the line frequency set by the camera, an image correction algorithm of the gear meshing surface has been proposed, which has been proven to have improved the accuracy of the detection results of gear contact fatigue pitting corrosion. The detection accuracy of fatigue pitting corrosion is improved by combining the preliminary detection and the accurate detection of the fatigue features. The depth information of the extracted contour pitting pits is extracted by the sequential forward selection (SFS) algorithm. The experimental results showed that 0.1216 m m 2 is the average absolute error of pitting corrosion detection, the average relative error is 2.2188%, and the detection accuracy is 97.7812%. The proposed pitting corrosion detection system advances in visualization, quantification, real-time monitoring, and failure judgment with a new, to the best of our knowledge, experimental approach for gear contact fatigue pitting corrosion detection.