High precision correction processing techniques for satellite imagery

Abstract

Remotely sensed data from satellites can provide very useful and important information in earth resources observations. However, satellite images usually include several kinds of distortions.
It is necessary for the correction processing of distorions. The correction methods of a geometic and a radiometric distortions for Multi-Spectral Scanner (MSS) imagery have already developed and become available.
But some problems in a Thematic Mapper (TM) and a High Resolution Visible (HRV) images not to stand out in an MSS imagery caused as follows;
(1) The discontinuity distortion peculiar to TM imagery occurs due to the bi-directional scanning. So this distortion cannot be modelized with the MSS geometric correction algorithm.
(2) The geometric and the blur correction processes can be composed of two filtering processes with an interpolation and blur correction capability. The about 90% of the total processing time for geometric correction spend in the interpolation processing. If the two filtering processes is used, the time is two times as much as the MSS imagery processing.
Therefore, I have proposed the geometric models consisting of separate geometric models of the bi-directional scans and frequency compoments and the geometric correction methods taking into account the correction of the blur.
Consequentry, the development of the high-precision modeling techniques for the discontinuity distortin correction and the high-speed filtering techniques in a geometric and a blur correction are necessary for an advanced hereafter sensors.
This paper discribes about the recent topics a geometric and a blur correction techniques for the whiskbroom and the pushbroom sensor imagery like TM and HRV.