There have been several examples in which both Synthetic Aperture Radar (SAR) and Hyperspectral Imaging (HSI) systems collected data in support of military operations (SMO). FOPEN (Foliage Penetration) radar has been used to penetrate tree canopies in order to detect objects. On the other hand, spectral differences between targets and backgrounds are used in HSI systems. Both SAR and HSI systems may suffer substantial false alarm and leakage rates due to respective background clutter. It is expected that a combined SAR and HSI system will greatly enhance the detection and identification performance. Based on the features derived from SAR and HSI data, a fusion approach has been established. Data sets of SAR and HSI over a common area from the Dixie data collection (May 1997 from Vicksberg, Mississippi) are used in this paper to demonstrate the fusion approach. The site contained several camouflage nets and vehicles. One of the vehicles was covered under a camouflage net. Target detection will be shown for each data set based on RCS (Radar Cross Section) and spectral features. In particular, a transformation of the spectral measurements into principal components was used to reduce the dimensionality of HSI data as well as to facilitate spectral feature extraction and material identification. SAR and HSI detections were subsequently combined via image coregistration. The fusion results showed that false detections in the SAR image were greatly reduced with background characterization of trees from HSI and target detections were confirmed with detection of camouflage nets and material identification of vehicle paints.
Date of creation, presentation, or exhibit
Department, Program, or Center
Chester F. Carlson Center for Imaging Science (COS)
Hsu, Sue May; Kerekes, John; Burke, Hsiao-Hua; and Crooks, Steve, "SAR and HSI data fusion for counter CC&D" (1999). Accessed from
RIT – Main Campus