【作者】 刘帅； 陈军；

【Author】 LIU Shuai1,2,CEHN Jun2 1School of Info-Physics and Geomatics Engineering, Central South University, Changsha , China, 410083, 2National Geomatics Center of China, Beijing ,China,100044

【机构】 中南大学信息物理工程学院； 国家基础地理信息中心；

【摘要】 基于序列图像的目标三维重建在计算机视觉与摄影测量学领域是一项非常重要的技术和研究热点。文章提出一种基于物体横向切片的可量测三维重建方法,利用平面单应变换和未标定的侧面多视图图像进行物体三维重建的工作,方法主要包括两个步骤,第一,在求取图像之间的单应变换时,通过参考物重构目标轮廓点从二维到三维的转换;第二,在求取物体的高度时,通过加入场景中的约束,提出一种快速、稳健的求高算法。文章采用的方法无需相机标定,无需求取基本矩阵,无需计算可视凸包,极大降低计算复杂性,试验部分以一个路牌重建为例子,并求出重建后各边的长度和高度与实际长度和高度进行比较,试验取得了良好的结果,使得普通数码相机在三维数据获取中非常的方便、快捷,极大降低生产成本与工作量。更多还原

【Abstract】 Three dimensional (3D) reconstruction considering of image sequences is very important technology, which lies an unresolved and active research topic on the cross-section of computer vision and digital photogrammetry. In this paper, we focus on 3D object reconstruction from uncalibrated images and put forward a rapid metric approach to recover the object based on slicing planes. The approach done comprises two key steps, for the one hand, calculate homography transformation to obtain every outline of the reconstructed object, for the other hand calculate the reconstructed object height by reference height proposed by the paper, and then recover the 3D object. This method requires no camera calibration, the estimation of the fundamental matrix or image visual hull; hence, it reduces the computational complexity. The experiment shows that the method is much validated and something useful is obtained.更多还原