Course Description
Mapping Land-Use Land-Cover Change (LULCC) with satellite remote sensing provides a reference for understanding the trajectories, patterns, drivers, and consequences of land-cover change. During the course, we will take a look at the advancement of classification and change detection techniques to map LULCC and land-use intensity. The course will cover state-of the-art non-parametric machine learning classification methods (e.g., SVM, Random Forests, hybrid classifications deep learning), enhancement of classifications, accuracy assessment, and image fusion techniques, and analysis of large data composites. We will primarily concentrate on utilization of freely available datasets: optical, SAR and LiDAR, such as Landsat and MODIS imagery, new products available via the Copernicus program, such as from Sentinel-1 and Sentinel-2 satellites, but also commercial datasets.
The course builds on prior knowledge of working with satellite imagery, such as passing via Bachelor level class-Introduction to Remote Sensing, Classification of Spatial Data. However, we will try to accommodate the students without prior experience in remote sensing. The course also complements the class on Remote Sensing of the Biogeosphere. The course will be particularly useful for students who envision interdisciplinary use of satellite remote sensing in the Biogeosphere and Anthroposphere studies (human dimensions of land-cover change).
We will concentrate on application of remote sensing to study land-cover transformation, such as urban sprawl, agricultural and forestry dynamics. However, students are more than welcome to bring their own research projects, as well as to suggest alternate topics in line with their own interest. Students are also highly encouraged to incorporate knowledge gained via other classes and to perform interdisciplinary projects.