Subspace-based Clutter Filtering for Improved SAR Target Detection

A.S. Paul and A.K. Shaw (USA)


Detection, Clutter modeling, Subspace Filtering, Foliage Penetration SAR, Ultra Wideband Radar, Physics based modeling.


: A new class of Subspace filter based algorithms is proposed for detecting targets in forest clutter. The training phase “learns” the clutter characteristics using local or global clutter subspaces. Both off-line and on-the-fly self-training versions of the algorithm are developed. These adaptive approaches utilize the Singular Value Decomposition (SVD) algorithm where small blocks of data in the neighborhood of a sliding test window are processed in real-time to estimate clutter characteristics. The clutter models are then used to nullify clutter in the test window. The proposed approach is effective for removing impulsive clutter and improves detection performance. An ultra-wideband (UWB) SAR simulation technique employing physical and statistical models is also presented [4, 9]. This joint physics-statistics based technique generates realistic images possessing “blob like” and “spiky” clutter characteristics of UWB radar data in forested regions. The proposed SAR image simulation technique avoids intensive computations required in low-frequency numerical electromagnetic simulation techniques and allows system designers to test in extended and realistic operating conditions. The proposed subspace filter-based detection algorithms have been applied on the simulated UWB data for performance comparison.

Important Links:

Go Back