Second-Order Statistical Texture Representation of Asphalt Pavement Distress Images Based on Local Binary Pattern in Spatial and Wavelet Domain

摘要: Assessment of pavement distresses is one the important parts management systems to adopt most effective road maintenance strategy. In last decade, extensive studies have been done develop automated for distress processing based on machine vision techniques. One structural components computer feature extraction method. application areas image processing, textural features provide more efficient information regions properties than other characteristics. this research, three different algorithms were used extract vector and statistically analyzing texture six various types asphalt surface distresses. The first algorithm images second-order statistics utilizing gray level co-occurrence matrix in spatial domain. second third algorithms, descriptors local binary patterns extracted wavelet transform domain, respectively. classification a combination K-nearest neighbor method Mahalanobis distance, indicates that two stages arranging levels edges by applying pattern (third algorithm) had superior result comparison with recognition separation Classification performance accuracy first, 61%, 75% 97%,