Abstract
The objective of this study is to propose the image quality correction method for the satellite multispectral image.
Generally speaking, the deconvolution with MTF (Modulation Transfer Function) such as inverse filter, wiener filter and so on, has been used so often for the satellite multi-spectral image quality correction. But as for the satellite multi-spectral image quality correction, there are some problems about their accuracy and processing time.
With the object of the simplicity of the algolithm and the adaptability to various kinds of satellite data, we proposed the RGS method (image Restoration method with Gauss-Seidel algolithm) for the image quality correction. The procedure of the correction is carried out by means of iterative tech-niques with Gauss-Seidel algolithm.
Furthermore the 1D-RGS method (one Dimensional RGS method) is extended to the 2D-RGS method. Care must be taken that the 2D-RGS method is treated as the pseudo deconvolution in the strict sense (Burger and Cittert, 1932).
In comparison with the restoration method using MTF, the results of this study are as follows :
1) It was found that the 2D-RGS method is one of the most superior method in image quality correction. Because, everyone can easily carry out the correction and get the same results.
2) It takes usually about an hour to make the color composite image with MTF, but only a few minutes with the 2D-RGS method and the operation efficiency is very high in image quality correction.
3) The 2D-RGS method can be easily carried out adjusting only two parameters called the instrumental function and the dumping factor.
4) As a result of applying the 2D-RGS method to MESSR, TM and HRV data respectively, it was found that the higher the resolution of satellite data, the more efficiently the 2D-RGS method can be carried out.
