Abstract: A reversible compression scheme for high-resolution remote sensing image is introduced.The method involves four steps.First, the whole image is divided into small blocks.Second, an integer wavelet transform is used to obtain a set of biorthogonal subclasses of images, and the image block is decomposed into a six-level pyramidal representation with Mallat algorithm.Third, an optimal bitallocation process is applied to computer the appropriate bit-rate for each subband.Finally, the wavelet subbands are encoded with a Bit-Plane Coding.It can be used to compress image at arbitrary bit rate.In the experiment, the authors select one aerial photo and four satellite images, from SPOT5, IKONOS, Landsat, China Earth Resource satellites, respectively, to test the performance of the algorithm.The experiment shows that the algorithm is as efficient as JPEG2000, despite its simplicity.