基于关系图的无人机影像三维重建

发布时间：2017-12-27 01:40

本文关键词：基于关系图的无人机影像三维重建　出处：《光学精密工程》2016年06期 　论文类型：期刊论文

【摘要】：针对增量式运动恢复结构算法在无人机影像三维重建中运行效率低的问题,提出了一种基于关系图的无人机影像批处理重建方法。利用无人机低精度的位置姿态测量系统数据概略估算影像间的重叠关系,在关联影像之间进行尺度不变特征变换匹配,使用随机抽取一致性算法和三视图约束进一步剔除弱关联的影像连接关系,再通过深度优先搜索得到最终参加重建的影像集连通分量,构建出强关联的稳健影像关系图。然后,根据影像相对位姿参数,由最小二乘法解算全局一致性旋转参数,利用三视几何约束和线性规划求解全局一致位置参数,最后进行一次光束法平差优化。实验结果表明:与经典的增量式重建Bundler算法相比,该方法运行速度至少提高了2.6倍,且生成的三维点数量增加了76.5%,更真实地恢复出场景的几何形态。该算法在改善重建效果的同时提高了处理效率,能够满足无人机快速响应应用的需求。
[Abstract]:Aiming at the low efficiency of incremental motion recovery algorithm in UAV 3D reconstruction, a batch reconstruction method based on relation graph is proposed. To estimate the overlapping relationship between images. By using the data of position and attitude measurement system in UAV low accuracy, by scale invariant feature transform in image matching relation between the use of random consistency algorithm and three view constraints to eliminate weak correlation image connection, through the depth first search to get the final image reconstruction in sets of connected components. Build a strong correlation diagram robust image. Then, according to the relative pose parameters of the image, the global consistency of the rotation parameters is solved by the least square method, and the global location parameters are solved by using the three geometric constraint and linear programming. Finally, the adjustment of primary beam adjustment is carried out. Experimental results show that compared with the classical incremental Bundler algorithm, the speed of the algorithm is increased by at least 2.6 times, and the number of generated 3D points is increased by 76.5%, which can restore the geometric shape of the scene more realistically. The algorithm improves the effectiveness of the reconstruction and improves the processing efficiency, and can meet the needs of the UAV rapid response application.
【作者单位】：信息工程大学;中国天绘卫星中心;
【基金】：国家自然科学基金(41471387,41301526) 地理信息工程国家重点实验室开放基金(SKLGIE2015-M-3-1)
【分类号】：TP391.41
【正文快照】： 1引言无人机遥感系统因具有灵活、机动、响应速度快等独特优点,而广泛应用在战场侦察、城市规划、资源勘查、林业普查等活动中[1]。目前无人机遥感已经成为卫星遥感、有人机遥感之外的一种重要遥感手段,如何由无人机获取的影像快速稳健地恢复场景的三维信息,已成为摄影测量和

【相似文献】