| 车载全景序列影像的位姿估计优化方法 |
| |
| 引用本文： | 张正鹏,张强.车载全景序列影像的位姿估计优化方法[J].测绘科学,2017(12):122-126. |
| |
| 作者姓名： | 张正鹏 张强 |
| |
| 作者单位： | 辽宁工程技术大学,辽宁阜新,123000 |
| |
| 摘 要： | 针对车载全景序列影像的位姿估计问题,该文提出了一种基于角元素和线元素的位姿估计优化方法。根据全景多视几何关系,建立了角元素和线元素多视约束的优化模型,并推导完成模型的位姿求解方法。然后作为初始外方位元素,加入稀疏GPS/INS的相机控制数据约束,采用全景球模型的稀疏光束法平差对位姿估计结果进一步优化。实验选择IP-S2车载移动测量系统获取的全景序列影像,对比分析单视图、全景视图的连续位姿估计、本文方法和GPS/INS控制数据。实验结果表明,该文方法可提高位姿估计的精度,并减少连续位姿估计产生的累计误差。
|
| 关 键 词： | 车载全景影像 多视约束 位姿估计 GPS/INS 光束法平差 |
A pose estimation optimization method of vehicle-borne panoramic image sequence |
| |
| Abstract: | A pose estimation method based on corner element and line element is proposed for pose estimation of vehicle-borne panorama image sequence.According to the multi-view geometric relation of the panorama image sequence,the pose estimation model of the angle element and the line element was established based multi-view constraint,and the method of solving the model was deduced.Then,as the initial elements of exterior orientation,the sparse GPS/INS camera control point constraint is added,and the sparse beam method adjustment of the panoramic sphere model was used to further optimize the pose estimation.The sphere panoramic images from vehicle-borne mobile surveying system were chosen so that a comparison analysis between the pose estimation of single camera and panoramic camera is carried out in detail.The results showed that the proposed method could improve the accuracy of pose estimation and reduce the cumulative error of the pose position. |
| |
| Keywords: | vehicle-borne panoramic images multi-view constraint pose estimation GPS/INS bundle adjustment |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
