Druhová klasifikace pater lužního lesa z multispektrálních UAV dat

Majer, Hugo Miroslav

Tree species classification of floodplain forest layers using multispectral UAV data Abstract Master's thesis deals with a tree species classification in a multi-layered floodplain forest using high spatial resolution multispectral UAV data. The main objective was to verify the possibility, develop a methodology, and conduct tree species classification of individual forest layers of the targeted forest area and to quantify the extent to which it is possible to map the forest understorey obscured by higher forest layers using optical data. The classification legend consisted of five tree species. The developed methodology utilizes a 3D photogrammetric point cloud derived from multispectral UAV data, from which the threshold heights of individual forest layers were determined. The determination of forest layer heights was based on visualizations of the point cloud at selected locations using profile perspective. Three forest layers were established: upper, middle, and lower. The point cloud was filtered into individual forest layers, and the resulting point clouds were grided onto a raster. Data collected during field surveys of study area were used to create training and validation datasets for each forest layer. The rasters of each forest layer were classified using an object-based classification and a...

host :: přihlásit Digitální repozitář
		Hledej		Nový záznam		Nápověda		O repozitáři