Detection of surface crack in building structures using 1d local binary pattern (LBP) algorithm and K-NN classifier

Abstract

The purpose of this project is to develop a building crack detection using the 1D-LBP algorithm and K-NN classifier. Surface cracks in building structures are treated as critical indicators of major structural problems and durability. The appearance of monolithic construction was also destroyed by the cracks. It takes a lot of time to detect the surface cracks manually. The way of detecting cracks manually is based on the experience of the person, and thus it is mainly a subjective judgment of the inspector. Therefore, automatic detection and classification of surface cracks is the highest priority task because it provides fast and reliable detection and analysis. There are a lot of feature extraction methods and classification methods for crack detection. Classic local binary pattern (LBP) is one of the most useful feature extraction methods. Moreover, the K-Nearest Neighbour (K-NN) classifier is a widely use classifier due to its simplicity. Due to the current methods in feature extraction are still improving, this project proposed a new characteristic extraction method to increase the performance of crack classification. In this project, the performance of a classification system with the one-dimensional local binary pattern algorithm (1D-LBP) and the K-Nearest Neighbour (K-NN) classifier. There are two stages in the classification system. Firstly, the 1DLBP algorithm will extract the normalized crack images features and save the data in a text file. Secondly, the K-NN classifier is used to classify the 1D-LBP based features from the first stage. There are two classes for the classifier to classify, which are positive crack versus negative crack and severe damage crack versus less severe damage crack. The classification performance is affected by the 1D-LBP based information and the value of K in the K-NN classifier.