Deep Learning for Detection of Pavement Distress using Nonideal Photographic Images

Abstract

In this paper, a deep learning approach for detecting pavement distress from nonideal photographic images of the road is investigated. Due to inconsistent data quality, part of the associated machine learning challenge is to produce training and validation data that bears coherent information sufficient for the task of successfully training a deep convolutional neural network that provides required detection performance. In the paper, the proposed method for detecting pavement distress is described. Work-in-progress experimental results are reported and analyzed.