On-machine image measurement of tool position by image completion using time sparsity

Abstract

By high resolution of machining, on-machine measurement and in-process measurement of machining tool position are needed. Image measurement is tried to use for on-machine measurement of machining tool position. However, when we use image measurement, machining tool is hidden in case that obstacles such as cutting chips and sparks are generated. To overcome this problem, we propose to use time sparsity of obstacles generation. We detect obstacle using time difference of pixel value and complement hidden area by machine learning.