콘크리트 구조물 균열 인식 알고리즘 개발

초록

In this research, the background image was primarily removed from the crack image of concrete structure by binarization processing, and then the matrix convolution was conducted by using row and column matrix. After the intersection calculation of these two images, specimen mark was extracted without noise. In this criteria, specimen shape was recovered from the binary image, and the x, y coordinates of the specimen were acquired by using morphological reconstruction. With the same method, crack was successfully recognized from the unique characteristics of shape and brightness.