Crack Detection for Concrete Surface Images Based on Likelihood Ratio Test and Elimination of Straight Line Obstacles

Abstract

This paper presents a crack detection algorithm for noisy images of concrete structures based on the likelihood ratio test and the elimination of straight line obstacles. In order to detect candidates for the cracks in the noisy image, we employ the likelihood ratio test using test function defined by the two-dimensional Gaussian function. The straight lines, e.g. conduit pipes and concrete connection, included in the detected candidates for the cracks are eliminated using the Hough transform.