Deep neural networks for terrain recognition task

This paper focuses on the problem of using artificial, deep neural networks in terrain recognition task based on data from vision sensor. Information about a terrain class is valuable for mobile robots, as it can improve their motion control algorithm performance through the use of information about surface properties. In this work RGB-D sensor was used for providing vision data, which comprise a depth map and infrared image in addition to the standard RGB data. Our own model of the artificial neural network is presented in this work. It was trained using the latest machine learning libraries. The results of this work demonstrate the performance of artificial neural networks in the terrain recognition task and give some hints how to improve classification in the future.