在彩色图象传输、采集、存贮过程中，一般都采用对色度通道进行压缩的技术来减少图象占用的系统资源，在重建中，采用大面积着色的方法来重建图象。但这种方法会在图象的细节部分引入较大的误差，为了在实际应用中得到最佳的重建彩色图象，就需要研究在色度通道亚采样的情况下，如何利用了完备的数据有效地重建图象。针对CCD彩色图象的空间量化特点，提出了一种利用自然图象的约象条件和人眼视觉生理学特点重建CCD彩色图象的方法。该方法利用高分辩率的通道数据来获得图象的细节信息，并且利用图象各通道间的细节相似性（同变性）来获得低分辩率通道的图象细节。通过实验证明，该方法有效地提高了图象的清晰渡，消除了CCD图象重建中常见的伪彩和图象模糊等现象，可以用于数码相机或多光谱遥感图象处理等方面。

To compress the data during image digitalizeing, storage and transmission, We always use subsample method to sample the color channels in order to cut off data rate, Then use bilinearity method to reconstruct it. How to find the best way to reconstruct the color information lead us to study the principles of subsample problem. The Nayquest law limits the signal band width to half of sample rate. In subsample condition, this point will reduce sharpness of image and cause false color on image edges. This paper is focus on the problem of image reproduction under subsample condition. On study of the features of nature image and human eyes, We found there are similarities in details of image RGB color Channels. After analyse the subsample problem, We found it's possible to rebuild the subsample image if we have the information of image grade. Take consideration of this two points, we bring forth a new way to build the CCD color image, and found this way will improve the image sharpness remarkably.