Data handling challenges for remote sensing systems

Multi-spectral imaging systems in remote sensing satellites generate large volumes of image data. Information content increases as instruments use wider swaths, more spectral bands, finer resolution and larger bit depths. As an example, the next generation Landsat system is expected to have an instrument data rate of several gigabits per second. Solid state recorders appear to be the best storage technology for most remote sensing missions, but cost and size become prohibitive for multi-terabyte systems. With downlink bandwidth representing a bottleneck in the system, real-time on-board processing, particularly image compression, is a key to using on-board storage efficiently and reducing downlink data volume. However, the inherent constraints of space applications - size, weight, and power-generally preclude the use of software-only solutions. Reconfigurable computing should be considered for its potential to enhance system flexibility and performance by allowing the data handling subsystem to be reprogrammed from the ground to incorporate mission changes and/or improved algorithms.