Image classification is a scheme of adaptive image coding

[1]

proposed a vector scalar classification for the ordinary image adaptive coding

which is characterized by both less encoded side information and highly efficient coding. However

we found that it is not apparently efficient for the original VSC to be applied to satellite image coding. According to the characteristics of both the affluent edge information and the high frequent component in satellite images

the novel strategy based on

[1]

for partitioning frequent component was proposed

which has improved clustering efficiency. The coding efficiency is justified clearly in the classification by our simulation experiment through C++ program. The new method can be widely applied to both the adaptive coding and high bit rate coder design for satellite images.