Auto-focusing Technology in Perpendicular Length Measurement of Video Measuring Machine

Abstract

The higher the auto-focusing accuracy is, the smaller the perpendicular length measurement error of video measuring machine would be. The improved Tenengrad gradient evaluation function is selected by comparing the performance indexes of definition evaluation function. The full window is selected as the focus window. A search strategy combining rough and fine focus is proposed. In rough focus, mountain climbing search algorithm combined with binary search algorithm is used. And Gaussian curve fitting method combined with global search algorithm is used in fine focus. The experimental results show that the auto-focusing algorithm has good robustness and high accuracy which can reach 5μm (1/20 depth of field), and the average focusing time is 10 seconds which can meet the real-time requirements.