面向RGB－D数据的4D－ICP点云配准方法

南京大学学报(自然科学版) ›› 2018, Vol. 54 ›› Issue (4): 829–.

面向RGB－D数据的4D－ICP点云配准方法

苏本跃1，2*，韩　韦1，2，彭玉升2，3，盛　敏2，3

出版日期:2018-04-30
作者简介:1.安庆师范大学计算机与信息学院，安庆，246133； 2.安徽省智能感知与计算重点实验室，安庆，246133； 3．安庆师范大学数学与计算科学学院，安庆，246133
基金资助:
基金项目：国家自然科学基金(61603003，11471093)，教育部科技发展中心“云数融合.科教创新”基金(2017A09116)，安徽省高校优秀拔尖人才培育资助项目(gxbjZD26) 收稿日期：2018－04－02 *通讯联系人，E－mail：bysu@aqnu.edu.cn

4D－ICP point cloud registration method for RGB－D data

Su Benyue1，2*，Han Wei1，2，Peng Yusheng2，3，Sheng Min2，3

Online:2018-04-30
About author:1.School of Computer and Information，Anqing Normal University，Anqing，246133，China； 2．The Key Laboratory of Intelligent Perception and Computing of Anhui Province，Anqing，246133，China； 3．School of Mathematics and Computational Science，Anqing Normal University，Anqing，246133，China

摘要/Abstract

摘要： 针对三维彩色物体的配准问题，提出一种面向RGB－D数据的点云配准方法. 首先利用主方向贴合方法将待配准的两片点云快速拉近，使它们近似对齐；在点云精确配准阶段，将RGB颜色值转换成单通道的灰度值，并将灰度值范围映射到几何数据的范围，由映射后的灰度值和点云的几何信息构成四维向量；然后由点的局部邻域几何信息和颜色信息构造混合特征描述子，根据混合特征描述子获得源点云的特征点，在四维向量空间，利用k近邻算法在目标点云中搜索对应点，以提高搜索效率；最后，定义了一种基于4D欧氏距离的ICP算法，通过4D－ICP迭代算法实现点云的精确配准. 实验结果表明，面向RGB－D数据的4D－ICP配准方法，能够快速有效地实现RGB－D点云模型的配准，并在配准精度和保持颜色纹理方面效果突出．

Abstract: Aiming at the registration problem of color objects，a point cloud registration method for RGB－D data is proposed. Firstly，two point clouds are quickly registered by principal component analysis algorithm to make them roughly aligned. In the fine registration phase，RGB color values are converted into single－channel gray values，and the range of gray values is mapped to the range of geometric data. Four－dimensional vectors are formed from the mapped gray values and the geometric information of the point cloud. Secondly，the mixed features descriptors are constructed by the local neighborhood geometric information and color information. The feature points are obtained according to the mixed features. The k nearest neighbor algorithm is used to search the corresponding points in the 4D vector space for improving the search efficiency. Finally，a 4D Euclidean distance－based ICP algorithm is defined，and the fine registration of the point clouds is achieved by 4D－ICP iterative process. The experimental results show that the 4D－ICP registration method for RGB－D data can effectively achieve the registration of the RGB－D point cloud model，and has outstanding performance in registration accuracy and color texture preservation.

苏本跃1，2*，韩　韦1，2，彭玉升2，3，盛　敏2，3. 面向RGB－D数据的4D－ICP点云配准方法[J]. 南京大学学报(自然科学版), 2018, 54(4): 829–.

Su Benyue1，2*，Han Wei1，2，Peng Yusheng2，3，Sheng Min2，3. 4D－ICP point cloud registration method for RGB－D data[J]. Journal of Nanjing University(Natural Sciences), 2018, 54(4): 829–.

参考文献

[1]　Medan G，Shamul N，Joskowicz L. Sparse 3D radon space rigid registration of CT scans：Method and validation study. IEEE Transactions on Medical Imaging，2017，36(2)：497－506. [2] Jin H，Wang X，Zhong Z C，et al. Robust 3D face modeling and reconstruction from frontal and side images. Computer Aided Geometric Design，2017，50：1－13. [3] Cirujeda P，Dicente Cid Y，Mateo X，et al. A 3D scene registration method via covariance descriptors and an evolutionary stable strategy game theory solver. International Journal of Computer Vision，2015，115(3)：306－329. [4] Basha T，Avidan S，Hornung A，et al. Structure and motion from scene registration ∥ Proceedings of 2012 IEEE Conference on Computer Vision and Pattern Recognition. Providence，RI，USA：IEEE，2012：1426－1433. [5] Aiger D，Mitra N J，Cohen－Or D. 4－points congruent sets for robust pairwise surface registration. ACM Transactions on Graphics(TOG)，2008，27(3)：Article ID：85. [6] Yang J L，Li H D，Campbell D，et al. Go－ICP：A globally optimal solution to 3D ICP point－set registration. IEEE Transactions on Pattern Analysis and Machine Intelligence，2016，38(11)：2241－2254. [7] Yang J Q，Cao Z G，Zhang Q. A fast and robust local descriptor for 3D point cloud registration. Information Sciences，2016，346－347：163－179. [8] Besl P J，McKay N D. A method for registration of 3－D shapes. IEEE Transactions on Pattern Analysis and Machine Intelligence，1992，14(2)：239－256. [9] 王　欣，张明明，于　晓等. 应用改进迭代最近点方法的点云数据配准. 光学精密工程，2012，20(9)：2068－2077.(Wang X，Zhang M M，Yu X，et al. Point cloud registration based on improved iterative closest point method. Optics and Precision Engineering，2012，20(9)：2068－2077.) [10] Serafin J，Grisetti G. NICP：Dense normal based point cloud registration ∥ Proceedings of 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems. Hamburg，Germany：IEEE，2015：742－749. [11] Zhou Q Y，Park J，Koltun V. Fast global registration ∥ Leibe B，Matas J，Sebe N，et al. Computer Vision － ECCV 2016. Springer Berlin Heidelberg，2016：766－782. [12] 韩宝昌，曹俊杰，苏志勋. 一种区域层次上的自动点云配准算法. 计算机辅助设计与图形学学报，2015，27(2)：313－319.(Han B C，Cao J J，Su Z X. Automatic point clouds registration based on regions. Journal of Computer－Aided Design & Computer Graphics，2015，27(2)：313－319.) [13] 杨　玲，谯舟三，陈玲玲等. 结合Procrustes分析法和ICP算法的PICP配准算法. 计算机辅助设计与图形学学报，2017，29(2)：337－343.(Yang L，Qiao Z S，Chen L L，et al. PICP registration method based on procrustes analysis and ICP algorithm. Journal of Computer－Aided Design & Computer Graphics，2017，29(2)：337－343.) [14] 戴静兰，陈志杨，叶修梓. ICP算法在点云配准中的应用. 中国图象图形学报，2007，12(3)：517－521.(Dai J L，Chen Z Y，Ye X Z. The application of ICP algorithm in point cloud alignment. Journal of Image and Graphics，2007，12(3)：517－521.) [15] Men H，Gebre B，Pochiraju K. Color point cloud registration with 4D ICP algorithm ∥ Proceedings of 2011 IEEE International Conference on Robotics and Automation. Shanghai，China：IEEE，2011：1511－1516. [16] 苏本跃，马金宇，彭玉升等. 面向RGBD深度数据的快速点云配准方法. 中国图象图形学报，2017，22(5)：643－655.(Su B Y，Ma J Y，Peng Y S，et al. Fast point cloud registration based on RGB－D data. Journal of Image and Graphics，2017，22(5)：643－655.) [17] Pauly M，Gross M，Kobbelt L P. Efficient simplification of point－sampled surfaces ∥ Proceedings of 2002 IEEE Visualization. Boston，MA，USA：IEEE，2002：163－170. [18] Faugeras O D，Hebert M. The representation，recognition，and locating of 3－D objects. International Journal of Robotics Research，1986，5(3)：27－52.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed