On-Board Satellite Implementation of Wavelet-Based Predictive Coding of Hyperspectral Images

We present an algorithm for lossy compression of hyperspectral images generated by the MODIS instrument. To greatly reduce the bit rate required to code images, we use linear prediction between the bands. Once the prediction is formed, it is subtracted from the incoming band, and the residual (difference image) is compressed using the Set Partitioning in Hierarchical Trees (SPIHT) algorithm. The proposed bit plane-synchronized closed loop prediction simultaneously achieves high compression ratio, high fidelity, and simple on-board implementation.