Introduction

Detection and measurement of pavement cracking is necessary for keeping a well maintained road network. One of the earliest form of road defect is cracking. If left untreated, cracks will form into a worse condition such as potholes. Thus, early detection of road crack will allow road authorities to fix the defects earlier. The earlier they fix the defects, the cheaper it is to maintain the road.

This project aims to create a semi-automated crack detection system. The system presented uses existing image processing techniques to segment cracking information from the images. The techniques used are seed selection, line growing and noise removal algorithms. When testing the system on 123 one mega pixels images, we achieved 81.3% of successful classification. The images are classified as those that contain crack and those that do not. Please click Here to see the results obtained.

Key-words: Road crack, semi-automated crack detection system, pavement crack, road defects