Improving Real-Time Gear Contact Fatigue Inspection with Machine Vision

A fast gear image correction algorithm is proposed for the real-time detection of gear pitting failure area in gear contact fatigue test. Based on the fact that gear pitting first appears near the pitch circle of the gear and expands to the root circle [1], the algorithm uses the pitch circle as the basis for the camera scanning line frequency calculation, and uses the pitch circle as the segmentation point to calculate the actual pixel value of the gear meshing surface, which leads to a reduction in the amount of computational parameters. The trigonometric function calculation method instead of the involute parametric equation curve product formula is used, which makes the calculation speed increase. Based on the principle of involute gear meshing and the principle of scanning camera shooting, the pixel value of gear meshing surface imaging is deduced, as well as the formula for calculating the pitting rate of gears based on the pixel value is proposed. The experimental results show that the average absolute error of the gear image fast correction algorithm proposed in this paper is 2.2868%, and the average relative error is 0.11059mm2, the algorithm is relative to the gear mesh surface correction algorithm [2] in the detection of the gear pitting area of the absolute error decreases by 0.1412%, and the relative error decreases by 0.0127mm2, and the computational complexity for the gear mesh surface correction algorithm 1/5 of that of the gear mesh surface correction algorithm, and finally combined with the pixel value-based gear pitting rate calculation formula, thus accurately detecting the time of gear pitting failure.