The authors present an overview of remote sensing image compression using wavelet transforms. They begin by describing the wavelet properties that are most important for image compression. In particular, they present a method to construct bi-orthogonal wavelets and their finite impulse response (FIR) filter banks. All of these FIR filter banks all have linear phase characteristics and the signal can be reconstructed exactly. Next, they expound on a novel hybrid scheme that uses bi-orthogonal wavelets and the adaptive differential pulse code modulation (ADPCM) algorithm for very low bit-rate satellite image compression.

Complicated SPOT images of city scenes that contain many high frequency details, such as building structures and roads, are used in this investigation. Using their hybrid wavelet/ADPCM compression algorithm, the data integrity of the satellite image was maintained with a peak signal-to-noise ratio (PSNR) of approximately 26 dB while achieving a compression ratio of 150:1.