摘要：AbstractPotholes, cracks and patches are some types of road surface distresses whose assessment is essential in India. In the current field practices, road distress data assessment is reported to be done through distress data collection and processing of the collected raw data. At present, distress data collection is increasingly being automated by using various imaging systems. However, analysis of the collected raw video clips for distress assessment is still predominantly being done manually. This is expensive, time consuming and slows down the road maintenance management. In this paper, a robust method for automated detection and assessment of potholes, cracks and patches from real life video clips of Indian highways is proposed. In the proposed method, potholes, cracks and patches are detected and quantified automatically using various image processing techniques supported by heuristically derived decision logic. For testing its performance, the proposed method has been implemented under a Windows environment using OpenCV library. The results are evaluated through accuracy and precision- recall metrics and compared with the methods presented by earlier researchers as well as current practices in the field. And the proposed method is found to be more robust and efficient. The information extracted using the proposed method can be used for determining maintenance levels of Indian roads and taking further appropriate actions for repair and rehabilitation.
