This study introduces a two-stage method for monitoring impervious surfaces (IS) with high accuracy and resolution. We apply U-Net CNN to mask stable areas and Continuous Change Detection (CCD) for IS change analysis over time. Achievements include excellent IS detection (OA=0.96), change identification (OA=0.94), and timing (OA=0.77). The approach emphasizes computational savings by targeting significant changes and demonstrates CCD’s efficacy using PlanetScope imagery for accurate IS tracking.
Speaker
Next up in State of the Map US
Previous talkAll the Places: Gathering Scraped Places Data for OpenStreetMap
All the Places is a project that uses open source web scrapers to capture data in a consistent, open format and make it easily accessible by anyone. In this talk, we’ll look...
