Neural Network Approaches in the Problems of Detecting and Classifying Roadway Defects

The purpose of this article is to develop an intelligent system for detecting and classifying roadway defects by means of neural network algorithms, as well as using stereo vision methods. The article describes a data collection method, a unique dataset collected, neural network algorithms for detecting and classifying roadway defects, a single digital platform on which all collected data is aggregated, as well as the principle of operation of the entire system. The developed algorithms have shown good results: the algorithm for detecting structured objects on the video image works with an accuracy of 96.78%, the algorithm for classifying defects of the roadway on the video image works with an accuracy of 92.01%.