全局一致性优化的无人机大倾角影像相对定向

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 o-rientation problem of iterative convergence and low robustness, a unmanned aerial vehicle relative orientation meth-od 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.