基于双目视觉的六旋翼无人机立体匹配算法

针对六旋翼无人机双目视觉成像时,经双目融合后反馈的图像噪点过多,以及图像精度不够理想的问题,提出了一种在匹配过程中融入全局差错能量最小化的区域立体视觉匹配算法。由于视差的求解是立体匹配过程中最重要的环节,因此本文利用最小化差错能量矩阵求解最优视差的原理。通过提高立体视觉的视差精度,从而减少视觉融合过程中因数据问题产生的噪声干扰,最终提高了对场景信息三维重构的准确度。通过分别在室内外的仿真实验与真实环境重构实验,验证了本文提出的基于双目视觉的六旋翼无人机立体匹配算法的有效性与可靠性。

Stereo matching algorithm of six rotor UAV based on binocular vision

During binocular vision imaging of six rotor UAV,there are more noisy points and poor image accuracy.Aiming at this problem,an area stereo vision matching algorithm is proposed,which integrate a global error energy minimization into the matching process.As parallax solution is the most important part in the process of stereo matching,the error energy matrix minimization is used to solve the optimal parallax.By improving the precision of stereo parallax,thus noise interference caused by data problems is reduced in the process of the visual fusion,eventually the accuracy of the three-dimensional reconstruction of scene information is improved.Through simulation experiments and real environment reconstruction in indoor and outdoor environment respectively,the validity and reliability of the stereo matching algorithm based on binocular vision of the six rotor UAV is verified.