基于双目视觉的无人机避障之研究  

Research on UAV Obstacle Avoidance Based on Binocular Vision

作　　者：王淏 潘峥嵘[1] 朱翔[1] WANG Hao;PAN Zheng-rong;ZHU Xiang(College of Electrical and Information Engineering,Lanzhou University of Technology,Lanzhou 730050,Chin)

机构地区：[1]兰州理工大学电气工程与信息工程学院,兰州730050

出　　处：《自动化与仪表》2018年第4期34-38,共5页Automation & Instrumentation

摘　　要：小型无人机快速避障是无人机研究领域的一个重要问题。基于其自身具有诸多不同于地面智能机器人避障的特点,实现快速避障较为困难,文中在双目视觉的基础上研究了无人机避障的方法,即从水平放置的2台摄像机所得图像中获取场景的深度信息,从而达到自主快速避障。The rapid obstacle avoidance of small UAVs is an important problem in the field of UAV research.It has many features that are different from the obstacle avoidance of intelligent robot on the ground. it is difficult to achieve the rapid obstacle avoidance of small UAV. Based on binocular vision research obstacle avoidance method of UAV,the depth information of the scene is obtained from the image obtained by the two cameras placed horizontally,to achieve autonomous obstacle avoidance quickly.

关 键 词：无人机 自主快速避障 双目视觉 深度信息 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置] V279[自动化与计算机技术—控制科学与工程]