There are seven processes that can be shows in Figure 1.
The first step is to convert all three RGB images into HSI
image. Then use these gray scale images to be the input image
for detecting the feature points. This step is very important
because if we cannot detect the strongly point cover all part of
the mountain’s images, the 3D point in the real world
coordinate cannot be complete for reconstruction the 3D
mountain surface. The accuracy of the 3D points in the real
world coordinate may be low. Then we should to determine the
algorithm that can specify the unique points detected from the
mountain surface. For accuracy assessment the existing DEM
will be used.