Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR) Parametric Study
Abstract
SAR ATR is a very complex problem that still has not been mastered. SAR ATR is difficult largely due to the fact that SAR imagery exhibits large variability. SAR imagery is a function of many variables called operating conditions (OCs) that can be subdivided into three large groups. The three main OCs are target, environment, and sensor. Sensor operating conditions deal with the properties of the sensor that have some of the largest effects on the formation of SAR images, including depression angle, squint angle, frequency, PRF, polarization, single/multi-look, sensor abnormalities, noise level, strip versus spot, and resolution. In the development and testing of SAR ATR algorithms to date the effects of sensor OCs have been given very little thought. The ultimate objective of this study is to develop a road map for studying various effects of varying sensor OCs on the performance of SAR ATR algorithms. For achieving this goal, we conducted literature searches to see how much had been done in sensor OC study. We also studied alternative data sources and the ways to generate SAR data related to the variation of sensor OCs to support SAR parametric study. In addition, we allocated and implemented a number of baseline ATR algorithms for the evaluation of their performance under the variation of sensor OCs. Our research has established an experimental paradigm for SAR parametric study.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 2003
- Accession Number
- ADA418766
Entities
People
- Kefu Xue
- Sam Sink
Organizations
- Wright State University