针对移动机器人依靠激光雷达感知环境的问题,提出一种基于分开-合并框架的直线特征提取算法。在分开阶段,用IEPF(iterative end point fit)算法对采集的激光雷达数据集合进行递归分割。在合并阶段,对由两相邻数据集合构成的新集合同时用EPF(end point fit)算法和总体最小二乘法进行拟合,如果两种拟合误差分别小于各自的阈值,则合并两个集合。合并阶段是一个递归过程,直到所有的两相邻数据集合都不满足上述合并条件才终止算法。对比实验结果表明,该算法大大降低了IEPF算法固定阈值所带来的过分割和欠分割的可能性,得到了很好的直线特征提取结果。
For the environmental perception of mobile robots with light detection and ranging(LIDAR), a method based on split-merge framework was presented herein. In splitting stage, a data set was segmented recursively with the IEPF(iterative end point fit) algorithm. In merging stage, both the EPF(end point fit) method and the total least square method were used to fit a new data set composed of the two adjacent segments. If their fitted errors were less than their thresholds respectively, the two data sets were merged into one. Merging stage was a recursive process for that it didn’t terminate until every two adjacent data sets didn’t satisfy the merging condition mentioned above. Results of comparative experiment show that the method presented reduces the probability of over-segmentation and under-segmentation resulted of the fixed threshold of the IEPF algorithm greatly and achieves good results of linear feature extraction.
Klimentjew D, Arli M, Zhang Jianwei. 3D Scene Reconstruction Based on a Moving 2D Laser Range Finder for Service- robots [C]//2009 IEEE Inter- national Conference on Robotics and Biomimetics. Piscataway, NJ: IEEE Computer Society, 2009: 1129-1134.
[2]
Ogaz M, Sandoval R, Chacon M. Data Processing from a Laser Range Finder Sensor for the Construc- tion of Geometric Maps of an Indoor Environment [C]//IEEE International Midwest Symposium on Circuits and Systems. Piscataway, NJ: IEEE Inc. ,2009: 306-313.
[3]
Borges G A, Aldon M J. Line Extraction in 2D Range Images for Mobile Robotics[J]. Intelligent & Robotic Systems, 2004, 40(3) : 267-297.
[4]
Borges G A, Aldon M J. A Split- and-merge Segmentation Algorithm for Line Extraction in 2D Range Images[C]//Proceedings 15th International Conference on Pattern Recognition. Los Alamitos, CA, USA: IEEE Computer Society, 2000:441-444.
[5]
Vandorpe J, Brussel H V, Xu H. Exact Dynamic Map Building for a Mobile Robot Using Geomet rical Primitives Produced by a 2D Range Finder [C]//Proceedings of the IEEE Int. Conf. on Ro boties and Automation. New York, NY, USA: IEEE, 1996: 901-908.
[6]
Viet N, Stefan G, Agostino M, et al. A Compari son of Line Extraction Algorithms Using 2D Range Data for Indoor Mobile Robotics[J]. Autonomous Robot, 2007, 23(2): 97-111.
[7]
Choi Y, Lee T, Oh S. A Line Feature Based SLAM with Low Grade Range Sensors Using Geometric Constraints and Aetive Exploration for Mobile Ro bot[J]. Autonomous Robots, 2008, 24(1): 13-27.
[8]
Pavlidis T, Horowitz S I.. Segmentation of Plane Curves[J]. IEEE Transactions on Computers, 1974, 23(8):860-870.
[9]
Fernandes L A F, Oliveira M M. Real time l.ine Detection through an Improved Hough Transform Voting Scheme[J]. Pattern Recognition, 2008, 41 (1) : 299-314.
[10]
Stefan G, Viet N, Roland S. Results on Range Image Segmentation for Service Robots[C]//Pro ceedings of the Fourth IEEE International Confer ence on Computer Vision Systems. New York, NY: IEEE Computer Society, 2006 : 53-60.
[11]
Chen Hongbo, Wang Qiang, Xu Xiaorong, et al. Line Detection in Traffic Sign Image Based on Im proved Hough transform[J]. Optics and Precision Engineering, 2009, 17(5) : 1111-1118.