利用多光谱传感器对同一观测对象在多个窄光谱范围上获得的图像，称为多光谱遥感图像。这类具有高空间和谱间分辨率的谱图像数据量大，其存储和传输都比较困难。因此，对海量数据进行有效的数据压缩便成了遥感资料应用中迫切需要解决的问题之一。为了防止有用信息的丢失而影响图像的进一步处理和应用，采用无损数据压缩方法是解决该问题的有效途径之一。本文分析了多光谱图像的空间和谱间相关性等特点，从多光谱图像预处理、预测和变换无损压缩方法等方面介绍了目前主要的多光谱图像无损压缩方法，并对其特点进行了对比分析。

Multispectral remote sensing images are those obtained from one object by the multispectral sensor in multiple narrow spectral ranges. Large amount of multispectral image data produced by high spatial and spectral resolution poses a great challenge to transmit and storage. As a result, it becomes imperative to develop an efficient compression technique. Since the loss of useful information will affect further processing and applications of the images, a lossless scheme is preferred. In this paper an analysis of the multispectral images space and spectral features is presented, and useful lossless compression techniques such as preprocessing, prediction and transformation are introduced, and also a comparative analysis of these techniques is discussed.