Rozpoznani objektu v LIDARovych datech

Abstract

The thesis explains the basic principles of artificial convolution neural networks. It describes problems of object detection from LiDAR data using machine learning. The next part deals with the description and comparison of available works solving given problem. The last (practical) part of this work is devoted to the description of implementation of the segmentation and classification method using the convolutional neural network using TensorFlow and NumPy libraries.