Registration for indoor RGB-D point clouds assisted by a quad- configuration target

doi:10.13474/j.cnki.11-2246.2020.0385

Abstract

Abstract: For 3D reconstruction of textureless indoor scenarios, we designed a quad-configuration target to assist pair-wise registration of indoor RGB-D point clouds. In detail, we first detected the quad-configuration target by filtering large curvature points with threshold. Then the target parameters and centers in adjacent point clouds were fitted by the random sample consensus (RANSAC) algorithm, matching the centers of the targets by fitting parameters, and achieving rough registration via rigid transformation with 4 points congruent sets. Next, we iteratively identified the overlapping areas between adjacent point clouds. Finally, we used the optimized overlapping area to fine-tune the rigid parameters between adjacent point clouds. In order to test the usability of our method, we used a Kinect camera to acquire 12 station point clouds for two different indoor scenes, respectively. Experimental results indicated that the root mean square error between adjacent point clouds is less than one point sampling interval, which demonstrated the robustness of the proposed method in indoor scene reconstruction with less texture.

Key words: indoor 3D reconstruction, textureless scene, Kinect, point cloud registration, quad-configuration target

CLC Number:

YANG Haijian, XU Ershuai, CHEN Wenjia, XU Zhihua. Registration for indoor RGB-D point clouds assisted by a quad- configuration target[J]. Bulletin of Surveying and Mapping, 2020, 0(12): 32-36.

References

[1] 周全. 双目视觉下的SLAM三维场景建图及物体识别研究[D].重庆:重庆大学,2018.
[2] 温秀兰,张腾飞,芮平,等.基于三维机器视觉的工业机器人定位系统设计[J].组合机床与自动化加工技术,2018(9):49-52.
[3] 王令文.结合点云数据与BIM技术的古建筑三维重建与信息化管理[J].测绘通报,2018(6):114-117.
[4] 张永庭,徐友宁,梁伟,等.基于无人机载LiDAR的采煤沉陷监测技术方法——以宁东煤矿基地马连台煤矿为例[J].地质通报,2018,37(12):2270-2277.
[5] 张勇,金学波.用单数码相机实现物体表面的三维重建[J].计算机工程与设计,2008,29(11):2926-2929.
[6] 张峰,史利民,孙凤梅,等.一种基于图像的室内大场景自动三维重建系统[J].自动化学报,2010,36(5):625-633.
[7] 邸凯昌, 万文辉, 赵红颖, 等. 视觉SLAM技术的进展与应用[J]. 测绘学报, 2018, 47(6):770-779.
[8] 危双丰,刘振彬,赵江洪,等.SLAM室内三维重建技术综述[J].测绘科学,2018,43(7):15-26.
[9] RUSU R B, BLODOW N, BEETZ M. Fast point feature histograms (FPFH) for 3D registration[C]//Robotics and Automation, 2009. ICRA'09. IEEE International Conference on.[S.l.]:IEEE, 2009.
[10] ZENG A, XIAO J. 3DMatch:learning local geometric descriptors from RGB-D Reconstructions[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).[S.l.]:IEEE, 2017.
[11] AIGER D, MITRA N J, COHEN-OR D. 4-points congruent sets for robust pairwise surface registration[J]. ACM Transactions on Graphics, 2008, 27(3):670-679.
[12] GE X M. Automatic markerless registration of point clouds with semantic-keypoint-based 4-points congruent sets[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2017, 130:344-357.
[13] BESL P J, MCKAY H D. A method for registration of 3-D shapes[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(2):239-256.
[14] RUSU R B, MARTON Z C, BLODOW N, et al. Towards 3D point cloud based object maps for household environments[J]. Robotics and Autonomous Systems, 2008, 56(11):927-941.
[15] ESTER M, KRIEGEL H P, SANDER J, et al. A density-based algorithm for discovering clusters in large spatial databases with noise[C]//Proceedings of the 2nd International Conference on Knowledge Discovery and Data Mining. Portland:AAAI, 1996.
[16] XU Z H, XU E S, WU L X, et al. Registration of terrestrial laser scanning surveys using terrain-invariant regions for measuring exploitative volumes over open-pit mines[J]. Remote Sensing, 2019, 11(6):606.