Thesis Details
Fusion of Radar and Visual Data for Remote Sensing
The aim of this thesis is to generate optical imagery in case of its unavailability. Optical and radar data from the past are used to generate such images. The main application field of this thesis is agriculture, where countless vegetation indexes can be used. In this thesis, for simplicity, only NDVI is utilized. Four datasets were created, each for the first three seasons of the year and the fourth that connects them all. As a solution for an image to image translation, Pix2Pix-cGAN was chosen. The results show the differences in the use of these datasets, between the different amounts and types of used pictures, as well as the interval adjustments between pictures. Our research found that the network is capable of creating plausible imagery with valid numerical values, but struggles to correctly utilize the information about the radar difference, which is important in order to evaluate plant development mainly when the optical imagery is unavailable. This thesis and its results are unique due to the geographically diverse dataset across Europe and the focus on agriculture, regardless of crop type.
GAN, Image to image transfer, Pix2Pix, Remote sensing, Earth observation, agriculture, dataset creation, radar imagery, optical imagery, NDVI, Sentinel 1, Sentinel 2, satellite imagery
Burget Lukáš, doc. Ing., Ph.D. (DCGM FIT BUT), člen
Češka Milan, doc. RNDr., Ph.D. (DITS FIT BUT), člen
Hradiš Michal, Ing., Ph.D. (DCGM FIT BUT), člen
Rozman Jaroslav, Ing., Ph.D. (DITS FIT BUT), člen
Zbořil František V., doc. Ing., CSc. (DITS FIT BUT), člen
@mastersthesis{FITMT25186, author = "Tom\'{a}\v{s} Strych", type = "Master's thesis", title = "Fusion of Radar and Visual Data for Remote Sensing", school = "Brno University of Technology, Faculty of Information Technology", year = 2022, location = "Brno, CZ", language = "english", url = "https://www.fit.vut.cz/study/thesis/25186/" }