针对大孔径静态干涉成像光谱仪(LASIS)的成像特点,提出了一种支持感兴趣区域(ROI)的非对称三维分层树集划分(AT-3DSPI HT)压缩算法。首先,对高光谱干涉图像进行非对称三维离散小波变换。然后,根据光谱信息的分布特点,采用ROI方法对不同区域的变换系数赋予不同的编码精度,以保护光谱信息。最后,采用改进的三维分层树集划分(3DSPI HT)算法编码高光谱干涉图像的小波变换系数。实验结果表明,该方法在8:1压缩比下,获得大于40 dB的峰值信噪比,同时有效地保护了光谱信息。 Aiming at the imaging characteristics of large aperture static interference imaging spectrometer (LASIS), an asymmetric three-dimensional hierarchical tree desegregation (AT-3DSPI HT) compression algorithm is proposed to support ROI. Firstly, the asymmetric three-dimensional discrete wavelet transform is applied to the hyperspectral interference image. Then, according to the distribution characteristics of spectral information, the ROI method is used to assign different coding precision to the transform coefficients of different regions to protect the spectral information. Finally, the wavelet transform coefficients of hyperspectral interference image are encoded by the improved 3DSPI algorithm. The experimental results show that the proposed method achieves a peak signal-to-noise ratio (SNR) greater than 40 dB at 8: 1 compression ratio while effectively protecting the spectral information.