Automatic estimation of bias parameters of RPC model and ground coordinates from multi-directional satellite images

Abstract

Advanced Land Observing Satellite (ALOS)/Panchromatic Remote-sensing Instrument for Stereo Mapping (PRISM) measured triplet images at forward, nadir and backward view directions, and the digital surface model (DSM) is generated from set of triplet images. Existing methods generate multiple DSMs from individual triplet images, and eliminate the effects of outliers by taking average. The proposed method in this paper solves multiple observation equations from all triplet images, and simultaneously determines bias parameters contaminated in the rational polynomial coefficient (RPC) model and ground coordinates. The experimental results showed that errors of the coordinates estimated by using multi-temporal triplet images were almost as small as those estimated with manually determined bias parameters in case of 4 or more sets of triplet images. As a result, we conclude that the proposed method is effective for stably generating accurate DSMs from multi-temporal triplet images.