Automated image processing technique for detecting and analysing concrete surface cracks

TLDR

The study proposes an image‑processing technique that automatically detects and analyses cracks in digital images of concrete surfaces. The technique employs morphological correction for non‑uniform brightness, enhanced binarisation and shape analysis, algorithms for width, length, orientation, and an artificial neural network to recognise crack patterns, and is implemented in a program validated against conventional methods. Experimental results demonstrate that the technique accurately measures and analyses crack characteristics.

Abstract

In the present work, an image processing technique that automatically detects and analyses cracks in the digital image of concrete surfaces is proposed. The image processing technique automates the measurement of crack characteristics including the width, length, orientation and crack pattern. In the proposed technique, a morphological technique was applied to correct the non-uniform brightness of the background, and enhanced binarisation and shape analysis were used to improve the detection performance; furthermore, detailed algorithms to calculate the crack width, length, orientation and an artificial neural network to recognise crack patterns including horizontal, vertical, diagonal (−45°), diagonal (+45°), and random cracks are proposed. An image processing program was developed for the proposed algorithm and a series of experimental and analytical investigations were performed to assess the validity of the algorithm. Then, the crack characteristics measured using the proposed technique were compared with those obtained using a conventional technique. The test results showed that the crack characteristics can be accurately measured and analysed using the proposed technique.