Next Generation of Image Analysis Methodology for Remote Sensing

The proposed SIP will be strongly linked to national and international collaboration within various current projects covering NR's three main application areas 1) monitoring environment and climate, 2) natural resource mapping and monitoring and 3) detect ion and mapping of man-made objects. The current activities in Area 1 are related to 1a) monitoring climate change in the cryosphere, 1b) detection of marine oil spill pollution and 1c) monitoring the state of forests and natural vegetation. For Area 2 th e activities are 2a) mapping snow for improved water management, 2b) forest resource mapping and 2c) mapping mountainous grazing land, and for Area 3 the activities are 3a) updating maps of new and changed infrastructure, 3b) vehicle detection for traffic statistics and 3c) detection of potential cultural heritage sites in agricultural fields. The proposed areas of methodological research and development are I) generic methodology for multi-sensor, multi-scale, multi-temporal and hyperspectral data, II) r etrieval of geophysical variables in the cryosphere, III) detection and classification of amorphous objects, IV) detection of man-made objects in very-high resolution imagery and V) high-accuracy pre-processing and lossless image compression. In synergy w ith NR's application-oriented projects, the SIP will contribute with key methodological frontier research results, while the other projects cover the more practical parts - including problem specification, establishing data sets, implementing large softwa re systems, validating and demonstrating the results, and finally the clients evaluating the results. When in-depth domain-specific expertise is needed - like botanic, forestry, glaciology and oceanography - NR's approach is to seek alliances with other r esearch institutes (nationally or internationally). Therefore, NR will also in this SIP work closely with other research institutes through the already established network of collaboration.