Target Discrimination in Polarimetric ISAR Data using Robust Feature Vectors
Abstract
We study the robustness of features against aspect variability and target modification for the purpose of target discrimination using polarimetric 35 Ghz ISAR data. The data are obtained with the MEMPHIS radar and comprise ISAR data of 16 targets providing imagery at a resolution of about 20 cm resolution for a complete aspect angle range of 360 degrees. The data cover three classes of military targets (T72, ZSU and BMP) with several modifications. For the study we have composed feature vectors out of individual radiometric, geometric and polarimetric features extracted from the imagery. Using the feature vectors and a nearest neighbor classifier we have determined how well different targets classes and different target modifications can be separated. We have found that good discrimination results are obtained between the target classes but that no discrimination is obtained between the different modifications.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 2005
- Accession Number
- ADA471110
Entities
People
- Bert Van Den Broek
- Rob Dekker