Authors: Luke Winters*, University of Denver
Topics: Remote Sensing
Keywords: Remote Sensing, Digital Globe, Planet, World-View, OBIA, Machine Learning
Session Type: Poster
Presentation File: No File Uploaded
The Mountain Area Land Trust (MALT), a non-profit organization which focuses on conservation of land in Colorado. For MALT, determining what water resources are located within their conservation areas is an important goal in the conservation efforts, and through this project, I am focusing on two of their priority conservation zones. I’m using satellite imagery from Digital Globe, Planet, and Sentinel 2 to test the performance of different datasets and classification methods to determine how much of a benefit proprietary data and algorithms provide over freely-available data and techniques. In addition to the satellite imagery, I am using other GIS data to aid in the classifications, including Digital Elevation Models and the National Wetlands Inventory. For processing, I am using Object-Based Image Analysis (OBIA) and implementing machine learning classification algorithms including Random Forest for classifications. For the coarser spatial resolution Sentinel-2 Images, I am comparing OBIA to pixel-based classifications, and machine learning algorithms to traditional classification methods, such as maximum likelihood and minimum distance.
To access contact information login