Abstract: Real-time obstacles detection of mobile robot in indoor environments based on the integrated system of stereo vision and laser range finder is investigated.Disparity information is obtained from stereo vision system.Fast method for obstacles detection is used for stereo vision system by fitting ground plane directly from disparity information;an algorithm named RANSAC（Random Sample Consensus） is used to filter the noise points and it can make the detection robustly.Then occupancy grid map is obtained from obstacles information on ground plane based on 3D world model.Grid data is extracted and then is used to fuse the integrated information provided by stereo vision system and laser range finder.Experimental results show that the integrated vision sensors of mobile robot not only obtain adequate obstacles information，but also ensure the real-time and accuracy performance.

Key words: sensors integration, disparity plane, Random Sample Consensus（RANSAC）, occupancy-grid, real-time obstacle detection

摘要： 研究移动机器人在室内环境下集成双目视觉和激光测距仪信息进行障碍物实时检测。由双目视觉系统检测环境获取视差信息，通过直接对视差信息进行地平面拟合的方法快速检测障碍物;拟合过程中采用了随机采样一致性估计算法去除干扰点的影响，提高了障碍物检测的鲁棒性。用栅格地图表示基于机器人坐标系的地平面障碍物信息并对栅格信息进行提取，最后把双目视觉与激光测距得到的栅格信息进行集成。实验表明，通过传感信息集成，移动机器人既得到了充分的障碍物信息，又保证了检测的实时性、准确性。

关键词: 传感集成, 视差平面, 随机采样一致性, 栅格, 实时避障