全局一致性优化的无人机大倾角影像相对定向  

Unmanned Aerial Vehicle Large Angle Image Relative Orientation Based on Global Consistency Optimization Method

作　　者：王一[1] 胡莘[2] 赵莹芝 

机构地区：[1]长安大学,陕西西安710054 [2]西安测绘研究所,陕西西安710054 [3]西安航天天绘数据技术有限公司,陕西西安710054

出　　处：《测绘科学技术学报》2017年第4期382-386,共5页Journal of Geomatics Science and Technology

基　　金：国家自然科学基金项目(4157012140)

摘　　要：针对无人机拍摄的大数据量大倾角影像三维重建,为获取影像的相对位姿,解决影像相对定向迭代不收敛、鲁棒性低等问题,提出了一种基于全局一致性优化的无人机相对定向方法。通过对增量式相对定向算法全局一次性平差优化改进,得到影像的高精度位姿信息。实验表明,该方法进行相对定向的效率与精度得到明显提高,并具有较好的鲁棒性。For the three-dimensional reconstruction of a large amount and large angle of unmanned aerial vehicle photograph data, in order to get the relative position and orientation of the image, and to solve the image relative orientation problem of iterative convergence and low robustness, a unmanned aerial vehicle relative orientation method is proposed based on global consistency optimization method. Through global one-time adjustment optimization to the incremental relative orientation, the high-precision position and orientation information of the image is obtained. Experiments show that the method is obvious to improve the efficiency and precision of relative orientation and has good robustness.

关 键 词：三维重建 相对定向 位姿解算 全局一致性优化 平差优化 

分 类 号：P231[天文地球—摄影测量与遥感]