Related Experiment Video
Updated: Jul 7, 2026
03:31
End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
Published on: December 15, 2023
Adaptive target detection in foliage-penetrating SAR images using alpha-stable models
A Banerjee1, P Burlina, R Chellappa
1Center for Automation Research, Department of Electrical Engineering, University of Maryland, College Park, MD 20742-3275, USA. banerjee@cfar.umd.edu
Summary
A new algorithm improves detecting targets in foliage-penetrating (FOPEN) ultra-wideband synthetic aperture radar (UWB SAR) images. It uses advanced clutter modeling and site context for better region-adaptive detection, overcoming limitations of conventional methods.
Area of Science:
- Radar Imaging
- Signal Processing
- Remote Sensing
Background:
- Detecting targets obscured by foliage in foliage-penetrating (FOPEN) ultra-wideband synthetic aperture radar (UWB SAR) imagery presents significant challenges.
- Conventional detection algorithms struggle due to differing target returns in foliage versus non-foliage regions and low signal-to-clutter ratios in UWB SAR data.
Purpose of the Study:
- To develop a novel algorithm for robust target detection in FOPEN UWB SAR images.
- To address the limitations of existing methods in accurately modeling clutter and utilizing local context for detection.
Main Methods:
- Incorporation of symmetric alpha-stable (SalphaS) distributions for precise clutter modeling.
- Construction of a two-dimensional (2-D) site model to derive local contextual information.
- Exploitation of the site model for region-adaptive target detection, including image segmentation and constant false alarm rate (CFAR) analysis.
Main Results:
- The proposed algorithm demonstrates improved performance in target detection within FOPEN UWB SAR images.
- The SalphaS distribution model is validated for its effectiveness in image segmentation and CFAR detection.
- Empirical results on real FOPEN images confirm the algorithm's robustness.
Conclusions:
- The novel algorithm offers a more effective approach to target detection in challenging FOPEN UWB SAR scenarios.
- Accurate clutter modeling using SalphaS distributions and leveraging local context via a 2-D site model are key to enhanced detection performance.