Thesis Details
Urban Element Detection Using Satellite Imagery
This thesis focuses on the right detection of objects in satellite imagery using convolutional neural networks. The goal of the thesis is to detect swimming pools and tennis courts in satellite imagery from different cities using the trained model. The model works with data from 10 different cities. The RetinaNet neural network model and Detectron2 library were used for development. The final trained model can detect objects with the average precision (AP50) at the level of 63.402 %. The thesis can be useful in the field of automating the acquisition of land surface statistics.
computer vision, object detection, image classification, instance segmentation, semantic segmentation, convolutional neural networks, RetinaNet, Detectron2
Burget Lukáš, doc. Ing., Ph.D. (DCGM FIT BUT), člen
Holík Lukáš, doc. Mgr., Ph.D. (DITS FIT BUT), člen
Martínek Tomáš, doc. Ing., Ph.D. (DCSY FIT BUT), člen
Matoušek Petr, doc. Ing., Ph.D., M.A. (DIFS FIT BUT), člen
@bachelorsthesis{FITBT23945, author = "D\'{a}vid Oravec", type = "Bachelor's thesis", title = "Urban Element Detection Using Satellite Imagery", school = "Brno University of Technology, Faculty of Information Technology", year = 2021, location = "Brno, CZ", language = "english", url = "https://www.fit.vut.cz/study/thesis/23945/" }