POTHOLE DETECTION IN ROAD USING IMAGE PROCESSING

Automatic detection and characterization of cracks in road surfaces, which is used to detect and characterize the type of cracks and find the severity level of cracks, used to reduce errors in manual calculation. Road pavement images are converted into non overlapping image blocks and then features of image blocks such as mean and standard deviation are obtained. Then connected component algorithm is applied to detect the crack blocks. Using Otsu thresholding, severity level of cracks is identified. To improve performance, connected component algorithm is applied to features of images.