Reduced Complexity Wavelet-Based Predictive Coding of Hyperspectral Images for FPGA Implementation

We present an algorithm for lossy compression of hyperspectral images for imple-mentation on field programmable gate arrays (FPGA). To greatly reduce the bit raterequired to code images, we use linear prediction between the bands to exploit thelarge amount of inter-band correlation. The prediction residual is compressed usingthe Set Partitioning in Hierarchical Trees algorithm. To reduce the complexity of thepredictive encoder, we propose a bit plane-synchronized closed loop predictor that doesnot require full decompression of a previous band at the encoder. The new techniqueachieves almost the same compression ratio as standard closed loop predictive codingand has a simpler on-board implementation.