基于摄影测量学和图像光流的三维重建方法

Method of 3D reconstruction based on photogrammetry and image optical flow

传统的获取物体三维(3D)数据的方法需要很多额外的硬件设备且不适合户外的大场景的三维重建。为此,提出了基于摄影测量学和图像光流的方法,只需要通过一台单反相机拍摄系列图片即可获得物体的三维数据,且适用于户外场景的三维数据的获取。根据近景摄影测量学的方法,能够获得相机的内外参数;根据光流法,能够获得两幅图片上的图像点匹配信息以及颜色信息,根据相机的标定数据、匹配数据以及颜色数据,获得带有颜色纹理信息的物体三维点云。这种测量方式只需手持相机对物体拍摄多幅图片即可完成三维重建,可以用于大场景虚拟现实(VR)中的三维获取。

The traditional methods to obtain three-dimensional（ 3 D） data need a lot of extra hardware device,and are not suitable for 3 D reconstruction on large scenes outdoor. The method based on photogrammetry and image optical flow is proposed,it only needs a camera to obtain for 3 D data by shooting a series of pictures,and it can be used for 3 D data acquisition in outdoor scene. According to the method of close shot photogrammetry,it can obtain intrinsic and external parameters of the camera; and according to the optical flow method,it can get image points matching information and color information on two images,according to camera calibration data,matching data and color data,it can obtain 3 D point cloud with color and texture information. This method can complete 3 D reconstruction only needs hand-hold camera to shoot multiple images of the object,it can be used in 3 D acquisition in large scenes virtual reality（ VR）.