Detekce vozidel v leteckých snímcích

Abstract

This bachelor's thesis is about the implementation of an application for vehicle detection in aerial images. This application can be used in traffic safety for traffic jams and accident detection or in assistance elements such as information about the availability of spots in parking lots or density of traffic. The final application is based on YOLO architecture which is based on a neural network and its ability to learn from training with a custom dataset. Bachelor's thesis besides application explains theory which is needed to understand the topic and finally a comparison of different architectures on the given problem.