Extraction of Building Features from Stand-Off Measured Through-Wall Radar Data

article
Automated extraction of building features is a great aid in synthesizing building maps from radar data. In this paper, a model-based method is described to detect and classify canonical scatters, such as corners and planar walls, inside a building. Once corners and walls have been located, a building map can be synthesized. To detect and classify the canonical scatterers, sparse reconstruction with an overcomplete dictionary is used. The dictionary is tuned to phase models that are specific to the different canonical scatterers, i.e., a linear phase change for walls and a quadratic phase change for corners. This is a potentially robust method for detecting canonical scatterers because the polynomial degree of the measured phase is preserved, even after propagation through a wall. Building-feature extraction results obtained with actually stand-off measured through-wall radar data are discussed
TNO Identifier
533661
Source
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 9(Januari), pp. 149-158.
Pages
149-158
Files
To receive the publication files, please send an e-mail request to TNO Repository.