Normalized Algorithm with Image Processing Methods for Estimation of Crack Length

Roads are one of the major transportation networks in our society and people are routinely used this to reach their destination. So, roads are continuously and highly utilized throughout the world. It was one of the reasons for arriving cracks and holes on the surface of the roads. One of the significant pavement distresses is cracking; provide the necessary information and guiding the authorities for road maintenance activities. It helps the concern authorities to maintain roads under good condition and improves safety status of the people. Human visual detection for distress of roads is traditional detection technique. It is time consuming, high expensive, labor-intensive and risky with high traffic density. In this work developed a simple image processing algorithm for estimation of crack length from ground truth image captured at a height of 4 feet with 16 mega pixel and 8 mega pixel mobile cameras. Histogram stretching was performed for the better visualition of the cracks, then convert it to gray image and resize to [256 256]. Select the end points of the crack in the resized image and determined the Euclidean distance between the end points, further it was normalized into centimeter scale and compared with ground truth value measured physically while captured particular crack image.