基于多摄像头的AGV全局导航地图创建

摘要

为了创建AGV大范围全局导航地图,论文提出一种基于多摄像头的全局导航地图创建方法.首先,在AGV活动区域上方垂直安装多个摄像头采集大范围区域的局部图像；其次,通过相位相关法和改进的SURF特征匹配相结合的算法对四幅局部图像进行拼接；最后,采用基于粒子群的模糊C均值聚类算法对全局图像进行分割提取障碍物信息,并建立室内环境全局导航地图.实验表明该方法与现有算法相比具有更好的实时性,能够快速建立全局导航地图.%In order to create wide-range of AGV global navigation map, a map building method for global navigation which is based on multi-camera is proposed. Firstly, multiple cameras are installed vertically above the AGV activity area to acquisition wide-range area of local images. Then, splice the four local images by the way of the phase correlation method and improved SURF feature matching algorithm. Finally, using Fuzzy C-Means clustering algorithm based on particle swarm for global image segmentation to extract obstacle information, and to establish a global navigation map of the indoor environment. The experimental results indicate that the proposed algorithm not only has better real-time compared with the existing algorithms but also can establish global navigation map effectively.