针对目前视觉定位方法大多基于地面语义特征(如车道线、停车线等)容易受到其他地面语义特征(如箭头、斑马线等)的影响,提出了一种基于地面快速鲁棒特征(SURF)点的全局定位方法.该方法首先在鸟瞰图中检测SURF点，结合高精度GPS构建地面SURF地图.然后在此基础上，使用迭代最近点算法，将在线检测结果与地图匹配获得车辆全局定位，并通过扩展卡尔曼滤波将定位结果与惯导和里程计数据进行融合，提高全局定位精度.实验结果表明，所提出的方法可获得分米级定位结果，能满足智能车的定位需求.
Global localization is essential for intelligent vehicles as navigating on the urban road. Generally, visual localization methods are based on semantic landmarks such as lanes and stop-lines which are easily interfered by other semantic landmarks, such as arrows and zebra crossing. To solve the problem, a new global localization method using ground speeded up robust features (SURF) is proposed in this paper. Firstly, SURF extracted from bird-eye view images are fused with high-precision GPS data to create a priori map. Then, SURF extracted online are matched with the map to estimate the global localization using the iterative closest point (ICP) algorithm. Finally, the global localization is fused with other sensors data by the extented Kalman filter (EKF) for better accuracy. Experiment results show that localization can reach decimeter-level accuracy, which can meet the demand for intelligent vehicles.
［1］CHEN C, IBANEZ-GUZMAN J, LE-MARCHAND O. Low-cost looselycoupled GPS/odometer fusion: A pattern recognition aided approach［C］//International Conference on Information Fusion. Cologne, Germany: IEEE, 2008: 1-6.
［2］HATA A Y, OSORIO F S, WOLF D F. Robust curb detection and vehicle localization in urban environments［C］//Intelligent Vehicles Symposium Proceedings. MI, USA: IEEE, 2014: 1257-1262.
［3］WONG D, DEGUCHI D, IDE I, et al. Single ca-mera vehicle localization using SURF scale and dyna-mic time warping［C］//Intelligent Vehicles Symposium Proceedings. MI, USA: IEEE, 2014: 681-686.
［4］DEKEL S, LEVI D, SLUTSKY M, et al. Monocular self localization in an urban environment using a prior based soft optimization robust estimation method［C］//Intelligent Vehicles Symposium (IV). Gotenburg, Sweden: IEEE, 2016: 1-8.
［5］CUI D, XUE J, DU S, et al. Real-time global localization of intelligent road vehicles in lane-level via lane marking detection and shape registration［C］//International Conference on Intelligent Robots and Systems. USA: IEEE, 2014: 4958-4964.
［6］TAO Z, BONNIFAIT P, FREMONT V, et al. Lane marking aided vehicle localization［C］//International IEEE Conference on Intelligent Transportation Systems. The Hague, Netherlands: IEEE, 2013: 1509-1515.
［7］ZABIHI S M, BEAUCHEMIN S S, MEDEIROS E A M D, et al. Lane-based vehicle localization in urban en-vironments［C］//IEEE International Conference on Vehicular Electronics and Safety. Yokohama, Japan: IEEE, 2015: 225-230.
［8］KIM D, KIM B, CHUNG T, et al. Lane-level localization using an AVM camera for an automated driving vehicle in urban environments［J］. IEEE/ASME Transactions on Mechatronics, 2017, 22(1): 280-290.
［9］BAY H, TUYTELAARS T, Gool L V. SURF: Speeded up robust features［C］//European Conference on Computer Vision. Berlin Heidelberg: Springer-Verlag, 2006: 404-417.
［10］BESL P J, MCKAY N D. Method for registration of 3-D shapes［J］. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2002, 14(2): 239-256.
［11］TINE L, BRUYNINCKX H, DE SCHUTTER J. Kalman filters for non-linear systems: A comparison of performance［J］. International Journal of Control, 2004, 77(7): 639-653.
