Adaptive compression of multisensor image data

The objective of this study is to develop and test a methodology for the design of multisensor image compression systems. Data collected from an airborne active (laser) and passive (thermal infrared) imaging system are analyzed and an efficient adaptive transform coding method is developed to exploit the characteristics of the source. A number of adaptive techniques that compensate for the nonstationary nature of the source are evaluated, and a novel model-based bit allocation and quantization strategy is adopted. The compression/reconstruction scheme is implemented in a massively parallel processor, and a large number of multisensor images are processed in order to define the rate-distortion performance of the scheme. The utility of the reconstructed imagery is also evaluated by examining the performance of multichannel target detection algorithms as a function of bit rate.