Length and width of low-light, concrete hairline crack detection and measurement using image processing method

Abstract

Cracks on the concrete surface indicate the degradation of the structure. So, we should be very careful with the maintenance of the concrete surface; otherwise, cracks on the surface may cause severe damage to the structure, resulting in fatal accidents. Manual inspection is the oldest method for crack inspection. But it is also laborious, inaccurate, and erroneous. Since the manual approach completely depends on the specialist’s knowledge and experience, automatic image-based crack detection and its measurements became its replacement. There are already so many methods, and research is going on to automatically detect and measure cracks. But detection and measurement of cracks become very difficult when the image has low light and a very small, widening crack that is a hairline crack. This paper tried to detect the hairline crack on a low-light concrete surface image using the already existing morphological algorithm. But for low-light images with a hairline crack, we did not get any faithful results. So, we proposed a new algorithm and successfully measured the same for low-light images with hairline cracks. The detected crack length and width have also been calculated, and the detection error is also mentioned herewith.