Road Pothole Detection Based on AlexNet for Autonomous Driving

With the frequent use of cars, it has been an increase in the occurrence of potholes and cracks on roads, and in the road danger coefficient as well as the cost of road maintenance. In terms of road detection, powerful and efficient road recognition algorithms are utilized to realize contactless detection solutions, before which, numerous scholars had conducted extensive research in road detection. However, most road detection algorithms are tailed to road vehicle detection, instead of road surface detection. This paper proposes a novel training dataset and employs Alexnet Network to accurately detect road surface conditions in different environments. The algorithm also uses several evaluation parameters to calculate the accuracy of the algorithm (expand). The goal of this paper is to develop an efficient algorithm for the detection and recognition of road conditions in various situations, provide reliable road conditions, and simultaneously offer effective environment perception for the automatic driving system of vehicles.