中文摘要：

针对传统混合视觉系统在图像特征匹配时耗时长、匹配精度低和重建复杂等缺点，利用一个RGB—D摄像机和一个全景摄像机搭建了相应的新型混合视觉系统，提出一种新型混合视觉系统中共同视场的3维重建方法。该方法首先利用新型混合视觉系统的对极几何关系确定两摄像机共同视场区域。其次采用SURF（Speed Up Robust Features）方法分别获得RGB图像和全景图像共同视场内的特征点，利用最近邻匹配法找出特征点匹配对。然后通过RGB—D图像的深度信息分析和系统的坐标变换，完成共同视场内特征点的3维重建。最后，通过实验验证了该重建方法的可行性和有效性。

英文摘要：

With the shortcomings of time consuming and low matching precision in image feature matching and the reconstruction of complicated of the traditional hybrid vision system. A new hybrid system is constructed by a RGB - D camera and anomni -directional camera. Then a 3D reconstruction method for the common view field in new hybrid vision system is proposed. Firstly, the epipolar geometry of this system is utilized to determine the common view filed of this system. Secondly, the feature points of the common view file and RGB image were extracted using SURF and the corresponding matching points were found using nearest neighbor method. Thirdly, 3D reconstruction of the matching points is realized by the depth information and coordinate transformation of the system. Finally, the feasibility and effectiveness of the algorithm is verified in experiment.