Absolute position estimation using IRS satellite images

摘要

This paper presents an absolute position estimation method using an Indian remote sensing (IRS) satellite images, which is a part of a position estimation system. The accumulated buffer (AB) matching method is proposed, in which a set of accumulator cells is employed for a fast edge-based matching between IRS satellite images and aerial images. The proposed algorithm is based on the accumulated buffer, which is a set of accumulator cells employed in the Hough transform widely used for robust detection of lines or curves. Thanks to the inherent robustness of the proposed algorithm, accurate registration can be accomplished between totally different image modalities. By computer simulations with two sets of real aerial image sequences, the performance of the AB matching method is analyzed and its effectiveness is shown in terms of the position error in the hybrid position estimation system. We found that the proposed AB matching method significantly reduces the average position error produced by the relative position estimation in the hybrid position estimation system.