Pre-hardware optimization of spacecraft image processing software algorithms and hardware implementation

摘要

Spacecraft telemetry rates and product complexity have steadily increased over the last decade presenting a problem for real-time processing by ground facilities. This paper proposes a solution to a related problem for the Geostationary Operational Environmental Spacecraft (GOES-8) image data processing and color picture generation (GOES-8 application). The proposed solution is based on a PC platform and synergy of optimized software algorithms and reconfigurable computing hardware (RC) technologies, such as FPGA and DSP. The solution involved porting the GOES-8 application from its Silicon Graphics Inc Workstation/UNIX platform, making minor platform specific changes to the GOES-8 application so that it runs on the PC, benchmarking the various code segments, and implementing the most computing intensive functions in hardware. After pre-hardware optimization steps in the PC environment, the necessity for RC hardware implementation for bottleneck code became more evident. The problem was solved beginning with the methodology described by T. Flatley (AIST-0132-0000, 1999), M. Figueiredo et al (IEEE Comp. Mag., pp. 115-118, 1999), and S. Kizhner (Proc. ION GPS'2000), and implementing a novel methodology for this application. The PC-RC interface bandwidth problem for the class of applications with moderate input-output data rates but large intermediate multi source data streams has been addressed and mitigated. This opens a new class of satellite image processing applications for bottleneck problem solution using RC technologies. The issue of a science algorithm level of abstraction necessary for RC hardware implementation is also described. Selected software functions already implemented in hardware were investigated for applicability in order to create a library of RC functions for ongoing work. A complete class of spacecraft image processing applications development using reconfigurable computing technology to meet real-time requirements, including methodology, performance results and comparison with the existing system, is described in this paper.