Wall Crack Detection based on Adaptive Double Threshold Greyscale Transform

Abstract

The construction industry is an important industry supporting social-economic development. Detecting cracks in walls could help to maintain the conditions of buildings. A wall crack detection algorithm based on adaptive double threshold greyscale transform is proposed. The MATLAB built-in 'graythresh' function is modified in order to have a flag bit to preliminarily selection of the greyscale transformation threshold where the grey transform threshold and the binarization threshold can be automatically adjusted based on the image processing effect using a preliminaryselected threshold. The algorithm enhances the crack information and weakens the background information by multiple morphological operations of removing isolated small area pixels. The MATLAB simulation results show that the proposed algorithm has an accuracy of 96.16%, time from inputting the image to the completion of the labelling is 3-15 seconds, which depends on the complexity of the crack.