Towards the recognition of individual trees in decimeter-resolution airborne millimeterwave SAR

This paper discusses the potentials and challenges of the recognition of individual tree patterns in image data acquired by decimeter-resolution airborne synthetic aperture radar (SAR) systems working in the millimeterwave domain. Due to the different characteristics of conventional optical imagery and side-looking SAR imagery, sophisticated processing strategies have to be developed, which need to be based on a thorough understanding of the radar imaging effects. Two preliminary workflows are proposed on the insights gained in first real-data experiments: While one is based on 2D image processing techniques applied to gray-scale intensity imagery, the other one employs interferometric or tomographic pre-processing in order to reconstruct point clouds from multi-aspect and multi-baseline data. The tree patterns are then further analyzed in the 3D point data.