A compression algorithm of asymmetric 3D set partitioning in hierarchical trees (AT-3DSPIHT) supporting region of interest (ROI) is proposed for imaging characteristics of large-aperture static imaging spectrometer (LASIS). Firstly,the hyperspectral image sequences are decomposed with asymmetric 3D discrete wavelet transform (3D-DWT). Secondly,different coding precisions are assigned to wavelet coefficients in different regions by the method of ROI to protect the hyperspectral information. Finally,the transformed images are encoded with the adapted 3D set partitioning in hierarchical trees (3DSPIHT) algorithm. The experimental results show that the peak signal-to-noise ratio (PSNR) is more than 40 dB at 8:1 compression rate,and the efficient protection of hyperspectral information is achieved.