基于投影矩阵的空间非合作目标三维重建

Non-cooperative space target 3D reconstruction based on projection matrix

针对从摄像机运动恢复结构算法(structure from motion algorithm,SFM)重建结果精度低以及面片重建算法(patchbased multiview stereo,PMVS)计算时间长的缺点,提出1种基于投影矩阵空间非合作目标三维重建算法。首先采用针孔模型对三维空间点到二维图像像素点投影关系进行理论推导,然后利用对极几何约束原理求解出旋转矩阵和平移矩阵;再由投影矩阵,解算三维空间点坐标;最后结合精密可控转台,搭载双目相机进行仿真实验。仿真实验结果表明,与SFM算法和PMVS对比,重建结果点云个数分别增加了22倍和9倍,所需时间分别减少了5s和97s;验证该算法的高可靠性、高重建精度和强实时性。

In video processing, the reconstruction precision of structure from motion （SFM） algorithm is low; meanwhile, the re-construction process of patch-based multiview stereo （PMVS） technology is too complex and time-consuming. tion algorithm based on projection matrix is proposed. Firstly, based on the principle of pinhole, 3D point and image pixels in theory. Then by utilizing the principle of geometric constraints, to solve the rotation matrix and translation matrix And then uses the projection matrix to solve the 3D point coordinates. Finally, combining with precision con-trolled turntable with binocular camera make a simulation experiment Compared with the SFM and PMVS, the number of recon-struction points resutt has been increased by 22 times and 9 times^ and the time has been reduced by 5 seconds and 97 seconds to verify l^he algorithm^ s strong real-time performance and high reliability and high reconstruction accuracy.