摘要
本文采用了一种基于AKAZE特征检测和PnP算法的单目视觉测量方法对相机的相对姿态进行解算,用于快速准确地确定空间中两个目标间的位姿关系.采集合作目标的模板图像,提取附加到合作目标上的4个特征点的像素坐标,利用AKAZE关键点对模板图像和待测图像进行匹配并计算映射矩阵,通过映射矩阵得到4个特征点在待测图像中的像素坐标,然后结合合作目标的尺寸信息求解基于4个共面特征点的PnP问题,解算相机与合作目标的相对位置.实验分析表明该方法计算的实时图像相机位姿与真实结果接近,验证了本文方法的有效性.
In this study,a monocular vision measurement method based on AKAZE feature detection and the PnP algorithm is used to calculate the relative attitude of a camera and quickly and accurately determine the attitude relationship between two objects in the space.Specifically,the template image of the cooperative target is collected,and the pixel coordinates of the four feature points attached to the cooperative target are extracted.Then,the template image and the image to be measured are matched by key points of AKAZE,and the mapping matrix is calculated.After that,the pixel coordinates of the four feature points in the image to be measured are obtained through the mapping matrix.Finally,given the size information of the cooperative target,the PnP problem based on the four coplanar feature points is solved,and the relative positions of the camera and the cooperative target are calculated.The experimental analysis shows that the real-time image camera attitude calculated by this method is close to the real result,which verifies the effectiveness of this method.
作者
剌小兆
刘新妹
殷俊龄
LA Xiao-Zhao;LIU Xin-Mei;YIN Jun-Ling(State Key Laboratory of Electronic Testing Technology,North University of China,Taiyuan 030051,China)
出处
《计算机系统应用》
2023年第1期310-315,共6页
Computer Systems & Applications
关键词
姿态解算
AKAZE算法
特征匹配
透视n点定位
attitude algorithm
AKAZE algorithm
feature matching
perspective-n-point(PnP)