中文摘要：

提出了一种基于全景视觉的移动机器人同步定位与地图创建（Omni-vSLAM）方法．该方法提取颜色区域作为视觉路标；在分析全景视觉成像原理和定位不确定性的基础上建立起系统的观测模型，定位出路标位置，进而通过扩展卡尔曼滤波算法（EKF）同步更新机器人位置和地图信息．实验结果证明了该方法在建立环境地图的同时可以有效地修正由里程计造成的累积定位误差．

英文摘要：

A method of omni-vision-based simultaneous localization and mapping （Omni-vSLAM） for mobile robots is introduced. The presented method extracts the color areas and uses the extracted information as the visual landmarks. The imaging principle and localization uncertainty of the omni-vision are analyzed to build the observation model and locate the landmarks. Then the robot poses and map information are updated synchronously with the help of extended Kalman filter （EKF）. The experiment results show that the presented method can build the environment map and correct the accumulated location error caused by the odometer simultaneously.