中文摘要：

在光照条件可变且存在电磁干扰的环境下，针对机器人室外导航任务，提出了一种基于全景近红外视觉和编码路标的自定位系统．通过近红外光源照明，利用全景视觉识别采用条形编码格式的路标，并利用扩展卡尔曼滤波算法（EKF）融合视觉数据和里程计数据，从而实现机器人白定位．实验证明，该方法消除了室外大范围导航时光照变化对机器人定位结果的影响．

英文摘要：

A self-localization system based on near-infrared vision and bar-coded landmarks for robot navigating in outdoor environment with variable light conditions and electromagnetic interference is presented. The near-infrared illuminator and omni-directional vision are used for recognizing bar-coded landmarks. Data from vision system and odometry are fused with an extended Kalman filter （EKF） to realize robot self-localization. The experiment result demonstrates that the proposed method eliminates the effect of light variations on robot localization in outdoor long-range navigation.