Search

Envisat Advanced Synthetic Aperture Radar (ASAR) dual-polarization data are shown to be effective for regional forest monitoring. To this scope, an automatic SAR image preprocessing procedure was developed using SRTM DEM and Landsat TM image for geocoding in rugged terrain and smooth terrain areas, respectively. An object-oriented forest and non-forest classif ication method was then proposed based on the HH (horizontal transmit and horizontal receive) to HV (horizontal transmit and vertical receive) polarization intensity ratio and HV images of ASAR data at single acquisition time in winter. The developed method was applied to forest and non-forest mapping in Northeast China. The overall accuracy, the user’s accuracy and the producer’s accuracy of forest were 83.7%, 85.6% and 75.7%, respectively. These results indicate that the proposed method is promising for operational forest mapping at regional scale.

The present study proposes a concept and a joint processing chain for integrating the ESA’s upcoming Sentinel 1-3 satellites to address the needs from a series of users. The framework considers the advantages of the single Sentinel systems in terms of the spatial, temporal and thematic detail for the generation of a global land cover dynamics and change (LCDC) product. In addition, a calibration and validation plan is discussed and open science issues to fully implement the product have been identified. In order to (i) address key user requirements and (ii) develop concepts and processing techniques for a global LCDC product that go beyond the current state-of-the-art, user requirements and available global state-of-the-art land cover efforts and retrieval approaches of similar nature and their shortcomings were considered.