基于优化三维重建技术的快速影像拼接

Rapid Image Splice Based on Optimized 3D Reconstruction Technique

以无人机代替卫星遥感实现三维重建技术的影像拼接是一种成本低、灵活度高的实现方式。室外无人机影像拼接通常存在深度相机对环境光照条件十分敏感等问题,以运动恢复结构(SFM)和多视角立体(MVS)技术结合构建拍摄区域的密集点云,再通过密集点云实现影像拼接的方法能够解决上述问题,但存在计算量较大和实时性较差的缺点。采用SURF特征描述子和最近邻匹配方法降低计算量,优化算法实时性能,同时提出一种增量式SFM流程中最优图像添加策略,提高光束平差法效率,并降低误差。实验结果表明,该方法能够在较短时间内获得较高精度的拼接影像,具有良好实时性。

Replacing satellite remote sensing with drones to realize the image stitching of 3D reconstruction technology is a kind of implementation with low cost and high flexibility.The technology of image stitching with drones in outdoor areas usually has the problem that the depth camera is very sensitive to ambient lighting conditions.This problem can be solved by using Structure from Motion(SFM)and multi-view stereo(MVS)to construct dense point clouds in the shooting area and using the dense point clouds to realize image stitching.Nevertheless,it has drawbacks of large calculations and poor real-time performances.The method of SURF feature descriptor and nearest neighbor matching is used to reduce the computation and optimize the real-time performance of the algorithm.At the same time,an optimal image addition strategy in the process of incremental SFM is proposed,which improves the efficiency of the bundle adjustment method and reduces the possibility of errors.Results of the experiments show that the method proposed in this paper can obtain high-precision stitched images in a short time,and has good real-time performance.