METHOD TO COUNT TRACK IMAGE BY USING CANNY EDGE DETECTION

摘要

The present invention relates to a method to count a track image by using canny edge detection, capable of counting nano-tracks by using an image processing technique. The method includes: a step of optimizing an image including inputted tracks; a step of correcting and detecting the tracks by binarizing the optimized image; a step of recovering a damaged track among the corrected tracks; and a step of counting the total number of tracks, including the number of overlapped tracks, by using size information of the recovered track and the number of the detected tracks. The step of recovering the damaged track includes: a track contour detection step of detecting a contour, corresponding to the edge of a track, and then converting the contour into a line segment of a predetermined length; a suitable point estimation step of distinguishing a segment line, satisfying a connection condition, from the converted segment line of the track; and a model prediction and recovery step of predicting and recovering a model of the track by applying random sample consensus (RANSAC) algorithm to the segment line satisfying the condition. The contour of the track is detected by using canny edge detection during the track contour detection step.;COPYRIGHT KIPO 2016