Visual Measurement Methods for Gear Geometric Parameters in Automotive Transmission

A geometric parameter system based on the principle of vision was developed for automotive transmission gears aiming to improve detection effi-ciency and measurement automation to address low au-tomation and efficiency of measurement of geometric parameters with traditional methods in automotive transmission gear manufacturing. First, the images of gears were obtained through an image acquisition plat-form, some of which were pre-processed to separate their background information. An improved caliper measurement method was developed to address the light interference of the gear surface in the measure-ment. By using this method, the Deriche algorithm was used to extract image edges at the sub-pixel level, and the edge point information was obtained with a caliper and input into the Huber algorithm for edge fitting, which calculated the pitch diameter, modules, and geometric parameters. Finally, a geometric parameter measurement system was built for automobile transmission gears, and the measurement results were displayed using C# program-ming. The proposed system, with an average detection time of 2.86s, met the on-site accuracy requirements with a measurement precision of 0.02mm compared to the image-measuring instrument used in the studies. Compared with manual inspection, this system improves the inspection efficiency and automation degree, and can meet the actual production site of 5 s per piece of the beat demand, applicable to the production site of the rapid inspection.