Automatic Patch Detection

The road maintenance is based on periodic surveys and the cost incurred in this is very high. Many developing countries found it difficult to track the patch and designing maintenance schedule become cumbersome. Therefore, there is a need for inexpensive methods for road surveys. We propose a method using low-cost sensors which can provide quality results with low cost. In our work we have detected the patch and analyzed on the visual characteristics like its closed contour and its texture with the surrounding intact pavement. We have designed a texture filter to extract the texture and to create mask for detecting the patch from the acquired image. We have used standard deviation for detection. We have also found the area of each patch.