Detekce obsazenosti parkovacích míst pomocí algoritmu strojového učení bez učitele

Abstract

This bachelor thesis is focused on detecting the occupancy of parking spaces from still images. Contains a description of editing images needed for image pre-processing, description of unsupervised machine learning algorithm, the operating principle of classifiers K-means and Expectation Maximization. Implementation of these segments were using OpenCV practically tested on real images park and finally describes the results.