Statewide Change Detection Using Multitemporal Remote Sensing Data.

摘要

The USDA Forest Service and California Department of Forestry and Fire Protection are collaborating on a large-area change detection program using Landsat Thematic Mapper satellite imagery. The program covers the state of California in five separate project areas on a five-year cycle. The goal is to implement a long-term, low cost and high quality monitoring program to identify trends in forest health, assess changes in vegetation extent and composition, and provide data for updating regional vegetation and fire perimeter maps. Landscape changes are detected through two phases. In Phase I a Kauth-Thomas transform is employed to create a landscape-level change map that identifies a continuum of change classes. In Phase II the causes of change are quantified using ancillary information and fieldwork. Project areas are divided by vegetation type for analysis. Causal information about change is collected qualitatively in hardwood rangelands and conifer forests. Field data collected in conifer forests is used to calculate crown closure loss using regression analysis.