Buildings Cracks Classification Using Zoning and Invariant Moment Features and Quadratic Discriminant Analysis Classifier

Abstract

Natural disasters such as earthquake often cause cracks in buildings and even demolish them. The cracked building must be assessed by an expert to determine whether the building is still suitable for use or not. The feasibility of a building is assessed based on the width, depth, and length of cracks in walls, beams, columns, and even the floor of the building. Only experienced experts can do such kind of task so that building assessment requires many structural engineering experts when an earthquake has happened. However, structural engineering experts are limited which able to do buildings assessment in the area affected. Therefore, the research based on a pattern recognition approach is conducted to classify cracks in buildings to be mild, moderate, or severe. It will be part of automatic building assessment based on the crack analysis. An alternative pattern recognition approach for classifying buildings cracks is a scheme based on zoning and shape features and Quadratic discriminant analysis (QDA) classifier. Based on the experimental results the proposed scheme gives reasonable achievement more than 80% of accuracy.