SBIR-STTR Award

Systran Federal Corporation (SFC), the sister-company of Systran Corp., which is a Products Development and Marketing Company specializing in real-time, high-performance networking, and Prof. Yuan Zheng of the Ohio State University (OSU), are proposing to design and develop wavelet algorithm(s) and hardware for compression/decompression of Synthetic Aperture Radar (SAR) signals. The novelty of the innovation is based on wavelet transforms which will be applied in two ways. The first way is to use wavelets as the transmitting pulse for increasing the resolution, and the second is to use the wavelet transform to treat the sensor data for compressing the volume of information. By using this wavelet approach one will be able to increase the resolution while reducing the amount of data which have to be communicated to the ground command stations. As part of the Phase I Program,. SFC and OSU will demonstrate the feasibility of the proposed innovation by implementing wavelet algorithms on an FPGA (Field Programmable Gate Arrays) board. In Phase II, SFC and OSU will design and develop 'VLSI chips' for wavelet implementation. Both SFC and OSU have the required expertise to successfully carry out this SBIR program.

Benefits:
Development of 'innovative' wavelets-based solution for SAR image processing will immensely benefit the DoD. Wavelet implementation will increase efficiency and accuracy, and reduce latency.

Keywords:
Wavelets, SAR, Algorithms, FPGA, VLSI, Hardware, Compression/Decompression

Current and future intelligence, surveillance, and reconnaissance platforms exploit advanced Synthetic Aperture Radar (SAR) systems to identify critical targets. These systems generate enormous amounts of data requiring a large storage capacity and high transmission bandwidth. Advanced compression techniques are necessary to reduce the demands on storage and transmission while maintaining high data quality for human and machine analysis. Systran Federal Corp. (SFC), the Ohio State University (OSU), and Lockheed-Martin Corp. (LMCO) propose to develop, demonstrate, and commercialize a set of wavelet-based compression techniques adapted for SAR images and raw SAR data. As with other types of data, compression of SAR images must strike a balance between maximum density and minimum information degradation. Current techniques, such as those using a Discrete Fourier Transform or Discrete Cosine Transform, do not provide acceptable results due to the generation of artifacts. The research conducted in Phase I shows that wavelet-based techniques provide low loss of information at high compression ratios. The SFC/OSU/LMCO team will develop wavelet-based algorithms for compressing magnitude-only images and complex SAR data, and implement these algorithms in field programmable gate arrays for efficient and reconfigurable processing. Extensive testing will ensure that information vital to human and automated analysis is not degraded