基于极线及共面约束条件的Kinect点云配准方法

发布时间：2018-01-18 01:12

本文关键词：基于极线及共面约束条件的Kinect点云配准方法　出处：《武汉大学学报(信息科学版)》2017年09期 　论文类型：期刊论文

更多相关文章： Kinect 极线约束 点面约束 四点共面

【摘要】：Kinect作为轻量级手持传感器,在室内场景恢复与模型重建中具有灵活、高效的特点。不同于大多数只基于彩色影像或只基于深度影像的重建算法,提出一种将彩色影像与深度影像相结合的点云配准算法并用于室内模型重建恢复,其过程包括相邻帧数据的配准与整体优化。在Kinect已被精确标定的基础上,将彩色影像匹配得到的同名点构成极线约束与深度图像迭代最近点配准的点到面约束相结合,以提高相邻帧数据配准算法的精度与鲁棒性。利用相邻4帧数据连续点共面约束,对相邻帧数据配准结果进行全局优化,以提高模型重建的精度。在理论分析基础上,通过实验验证了该算法在Kinect Fusion无法实现追踪、建模的场景中鲁棒性依然较好,点云配准及建模精度符合Kinect观测精度。
[Abstract]:As a lightweight hand-held sensor, Kinect has the characteristics of flexibility and efficiency in indoor scene restoration and model reconstruction, which is different from most reconstruction algorithms based on color image or only depth image. A point cloud registration algorithm combining color image with depth image is proposed and used for indoor model reconstruction and restoration. The process includes the registration of adjacent frame data and the overall optimization, based on the Kinect has been accurately calibrated. The color-image matching points of the same name are combined with the point-to-surface constraints of the iterative nearest point registration of the depth image. In order to improve the accuracy and robustness of the adjacent frame data registration algorithm, the global optimization of the adjacent frame data registration results is carried out by using the contiguous constraints of the contiguous points of the adjacent frame data. In order to improve the accuracy of model reconstruction. On the basis of theoretical analysis, the experimental results show that the algorithm can not be tracked in Kinect Fusion, and the robustness of the modeling scene is still good. The accuracy of point cloud registration and modeling accords with the Kinect observation accuracy.
【作者单位】：同济大学测绘与地理信息学院;
【基金】：上海市自然科学基金(13ZR1444300)~~
【分类号】：TP212;TP391.41
【正文快照】： 随着信息技术的发展,三维模型重建技术受到各行各业的青睐,利用深度传感器获取深度数据(点云)进行三维建模及场景恢复是其中常用且重要的一种方法。三维激光扫描装置是常见的获取点云信息的传感器,但其价格昂贵,限制了点云数据在日常领域的应用。微软推出的一种彩色与深度传感

【相似文献】