A Novel Surface Crack Detection and Dimension Estimation Using Image Processing Technique<bold> </bold>

The authors present a novel crack detection system based on image processing. The authors have proposed a novel algorithm for detecting the crack on the surface and estimating the dimension of the crack. Monitoring the health of any character is essential, and the cracks developed on the surfaces often lead to a reduction in the strength of the material. Manual inspection is one of the most common methods for detection of crack. This approach is very time consuming and also depends upon knowledge as well as experience. It also lacks objectivity in the quantitative analysis. The proposed crack detection and dimension estimation using an image processing' system focus on automating the process using established digital image processing techniques. It is a standalone embedded device that detects the surface cracks through images captured via camera and estimates the size of the crack with proper calibration. The experimental results show that the mean error is 8.12%, and the accuracy achieved in dimension estimation of the crack is 91.88%.<bold> </bold>