Automated Detection and Identification of Area-based Distress in Concrete Pavements

Automated inspection and identification of area-based distress on concrete pavement surface is an important component of the pavement surface data collection and processing system. In this paper, we present an image processing method to implement high-speed, real time detections of various types of area-based distress, such as spalled cracks, punchouts, asphalt or concrete patches, in concrete pavements. In the algorithm, a statistical second order moment method is applied on a serial of multi-scale images to enhance the feature of distress region, followed by a region-based segment operation to discriminate the potential distress from the image background. By computing and estimating a set of feature parameters or properties, the extracted objects are further identified as a definite type of distress. Because the primary processing is performed on the coarse scale image, this algorithm also exhibits excellent computation efficiency. Real pavement images were processed using the developed algorithm, and the results demonstrated its validity for detecting area-based distress in concrete pavements