Impact of On-Orbit Data Compression on Signal-to-Noise Ratio for Dim Target Detection

摘要

Our capacity to transmit data off of a spacecraft to the ground station has not kept pace with rate of data generation from advanced sensors. One solution is to reduce the size of the data prior to transmission through advanced compression techniques. In this paper, software and hardware implementations of a novel data compression algorithm were used to compress overhead persistent infrared data with synthetic embedded dim targets. An image processing technique was subsequently applied to the decompressed data to recover the embedded targets within the image stack, and the signal-to-noise ratio of the detected targets were examined as a function of compression ratio. This paper examines the detection signal-to-noise ratio and relative speed-up for compression of 1200 frames of infrared data implemented in a general-purpose processor and a field-programmable gate array. The advantages of the FPGA implementation for processing efficiency were clearly shown, however, the reduced data precision had a negative impact on our ability to recover the dim target from the post compressed data.