摘要
随着人工智能的发展,在智能机器人及无人驾驶汽车应用上,空间定位是最基本和关键的问题。针对空间运动载体的定位算法,提出了双目视觉里程计定位算法,从摄像机标定、图像处理、特征提取和匹配以及运动估计4个方面展开了研究。出于实时性和精确性的考虑,使用Matlab2016b和Visual Studio2015两个软件进行混合编程。运用该算法对搭载双目相机的汽车在行进过程中所拍摄的图片进行计算,得出汽车精确的运动轨迹。
With the development of artificial intelligence,the spatial location is the most basic and key problem in the application of intelligent robot and driverless car.Aiming at the location algorithms of space moving carrier,a binocular vision odometer location algorithm is proposed.The research is performed on such four aspects as camera calibration,image processing,feature extraction and matching as well as motion estimation.Considering the real-time and accuracy,a complete theoretical system of binocular vision is formed by using the mixed programming of Matlab 2016b and Visual Studio 2015.This algorithm is used to calculate the image taken by the car with binocular camera in the process of moving and obtain the precise trajectory of the car.
作者
黄秀珍
王佳斌
高发钦
冉宇瑶
HUANG Xiuzhen;WANG Jiabin;GAO Faqin;RAN Yuyao(Keyi College,Zhejiang Sci-Tech University,Shangyu 312369,China;Zhejiang Sci-tech University,Hangzhou 315700,China)
出处
《无线电通信技术》
2019年第6期676-681,共6页
Radio Communications Technology
基金
2016应用技术型试点专业建设项目——计算机科学与技术(kyiszxzy1603)
浙江理工大学科技与艺术学院计算机科学与技术浙江省一流学科B类经费资助
关键词
视觉里程计
双目视觉
运动估计
图像处理
visual odometer
binocular vision
motion estimation
image processing
