METHOD FOR DETECTING PATTERN BLOCKS IN MACHINE VISION

摘要

PURPOSE: A method for detecting pattern blocks in a machine vision is provided to accurately and stably discriminate examination pattern blocks through gradient graphs in consecutive examination. CONSTITUTION: A method for detecting pattern blocks in a machine vision includes the steps of obtaining image information of the whole standard examination substance(S501); generating a gradient graph by each scan line(S503); setting a specific part for determining a one block by referring to the gradient graphs(S505); calculating distance from the corresponding specific part to both ends of an examination pattern block(S507); storing the calculation result of the image information of the graph corresponding to the specific part(S509); line-scanning an examination substance(S601); operating the sum of bright deviation of a plurality of cells, and obtaining a gradient graph of each scan line(S603); comparing the gradient graph of the present line scan with the stored gradient graph of the standard examination substance(S605); and obtaining a start point and an end point of the examination pattern block of the gradient graph of the present line scan(S607).