Abstract: Based on the principle of the Rate-Distortion slope ROI (Region of Interest) coding and the conventional bitplane
shift method, a new adaptive random shape ROI coding compression method for military satellite image systems is proposed,
which utilizes adaptive coding schemes for the low-frequency and the high-frequency components of the wavelet factors
respectively in terms of the size and shape of ROI. This adaptive method not only protects the information of a certain
object image but also obtains a high compression rate. Finally, experimental results show that when the compression rate is
16 or 32, the method satisfies the requirement for the military satellite remote sensing image system.

Key words: satellite remote sensing, image compression, ROI, bitplane shift, the rate-distortion slope