干旱区生态环境知识资源中心

A novel method is introduced for self-calibration and elimination of systematic errors for a position and orientation system (POS). The method uses a combined bundle block adjustment with POS data (named the POS-supported bundle block adjustment) without a calibration field. On the basis of delivering strict observation equations for POS data, the specific scheme of compensating the translation and drift systematic errors in a POS is given, and a prototype system WuCAPS is developed. The effects of eliminating POS systematic errors using the POS-supported bundle block adjustment for different ground control conditions are tested using two sets of actual aerial photos. The first set was taken over a flat region in the suburbs of the city of Yingkou in China and tested at a scale of 1:2500. The second set was taken over a high mountainous region in the desert region of Xinjiang in China and tested at a scale of 1:32,000. The empirical results verified that the POS systematic errors can be completely eliminated and the photo elements of exterior orientation obtained by the POS-supported bundle block adjustment can satisfy the requirements of aerial photogrammetric topographic mapping when four full ground control points (GCPs) are emplaced in the corners of the adjustment block for large-scale images of flat regions, but only one full GCP emplaced in the center of the adjustment block is needed for medium-small scale images of mountainous regions. This not only demonstrates the validity of the established mathematical model and the feasibility of the method proposed in this paper, but also avoids the use of a special calibration field. Therefore, it can simplify the existing POS operation rules and dramatically save on practical application costs, laying the theoretical foundation for widespread use of POS. (C) 2008 National Natural Science Foundation of China and Chinese Academy of Sciences. Published by Elsevier Limited and Science in China Press. All rights reserved.