Image Processing System for Automatic Scraper

摘要

An Image Processing System (IPS) for an automatic Surface Plate (SP) scraper system has been studied. First, the SP was took a red lead print, then the IPS made two images of the SP using a monochromatic ITV camera. One image was photographed through a blue filter glass, another was photographed through an orange filter glass. Secondly, values of densities of the image was subtracted from another. And the subtracted image was binalized at a threshold level which was determined by the new preprocess and OTSU method. Thus the red lead area was separated with a base area of the SP. Thirdly, the IPS calculated positions to be scrape in consideration of characteristics of a scrape unit and a robot control system.Before introduction of the IPS, it took 3 days to determine the scrape positions, but now it takes only 20min (max). As the IPS automatically processes all jobs except photographing, a beginner can treat the IPS.