Automated system for concrete damage classification identification using the discriminant analysis classifier

Abstract

Defects in concrete structures are major measures of structural integrity and serviceability. In association with the condition survey of concrete showing defects, a visual inspection is commonly used. However, this method is subjective, laborious, time-consuming, and complicated by demanding access to many components of large project design. Therefore, an automated system for concrete defects classification using the Discriminant Analysis Classifier has been introduced to obtain crack information more accurately from images. The objective of this research is to increase the efficiency of analyzing concrete defects in terms of quality, time, and cost. 200 images have been collected with 50 images data for each concrete defect which is crack, corrosion, spalling and non-defect for control data. The Gray Level Co- Occurrence Matrix (GLCM) is being used to produce an algorithm for image processing and feature extraction. This model is trained by using 80% of the images data and tested using another 20% of the image data. The result shows that the time reduced from hours to seconds in analyzing the concrete defect. Thus, the model achieved accuracy to 95% using the training data and 70% using testing data. This invention is very significant to helps engineers or construction inspectors during inspection activities.