基于激光雷达和Kinect信息融合的导盲机器人SLAM研究

SLAM research of seeing eyes robot based on Lidar and Kinect data fusion

针对室内复杂的未知环境,在采用激光雷达进行即时定位和地图构建SLAM的基础上,提出了一种融合三维深度信息和激光信息的导盲机器人SLAM方法.首先对比分析ROS机器人操作系统中几种主要的SLAM方法,通过深度传感器Kinect采集的信息,得到三维SLAM地图.为了弥补激光雷达采集数据方面的不足,先使用经典的Canny算法对地图进行边缘检测,再进行扩展相位相关计算得出地图间的平移、旋转和尺度变化,使用所得参数拼接地图,实现对三维SLAM地图和二维SLAM地图的融合.通过真实环境中的实验验证,该模型地图能足够全面和精确的识别环境和障碍物,机器人可以更好地完成导盲任务.

To deal with the complicated unknown indoor environment,a SLAM（ simultaneous localization and mapping） algorithm based on lidar data is proposed,which can fuse lidar data to Kinect data and be used in seeing-eye robots. Firstly several cardinal SLAM methods are compared and analyzed in Robot Operating System（ ROS）,then a 3 D SLAM map is created with the Kinect information. To make up the shortcoming of lidar on data acquisition,classical Canny edge detection method is introduced to detect map edge. Then the map translation,rotation and scale changes are calculated by the extended phase correlation method. These parameters are utilized to stitch 3 D SLAM map and 2 D SLAM map. Through the experiment in the specific real environment,the map can comprehensively and accurately recognize the environment and obstacles,and the robot can complete the task of blind guiding better.