基于航天应用的多圆特征识别和姿态估计
发布时间：2019-01-13 20:50
【摘要】：为了实现多圆特征识别和姿态估计,将航天器常见的几何特征如旋转体(SOR)应用于姿态估计,提出了一种基于椭圆归类的单目视觉姿态估计方法。在图像中采用基于弧段的椭圆检测方法检测目标上的椭圆特征;提出一种基于SOR空间圆平行性和垂直性约束的椭圆归类方法,得到合理的椭圆特征;利用这些特征估计航天器和摄像机之间的姿态。实验结果表明:该方法具有较好的椭圆归类效果和姿态估计精度,对于含有0~16%的椒盐噪声的仿真图,归类精确率不低于97%;实物实验中,角度误差不超过1°,深度方向(小于10 m)的测量误差不超过80mm,其他方向的测量误差不超过15mm。
[Abstract]:In order to realize multi-circle feature recognition and attitude estimation, a monocular visual attitude estimation method based on ellipse classification is proposed by applying the common geometric features of spacecraft, such as rotating body (SOR), to attitude estimation. The ellipse feature of the target is detected by using the ellipse detection method based on arc segment, and an ellipse classification method based on SOR space circle parallelism and perpendicularity constraint is proposed to obtain the reasonable elliptic feature. These features are used to estimate the attitude between the spacecraft and the camera. The experimental results show that this method has good ellipse classification effect and attitude estimation accuracy, and the accuracy rate of classification is not less than 97 for the simulation diagram containing 0 ~ 16% salt and pepper noise. In the experiment, the angle error is not more than 1 掳, the depth direction is less than 10 m, the measurement error is not more than 80 mm, and the other direction measurement error is not more than 15 mm.
【作者单位】： 哈尔滨工业大学电气工程及自动化学院;
【基金】：国家自然科学基金(51075095)
【分类号】：TP391.41;V448
[Abstract]:In order to realize multi-circle feature recognition and attitude estimation, a monocular visual attitude estimation method based on ellipse classification is proposed by applying the common geometric features of spacecraft, such as rotating body (SOR), to attitude estimation. The ellipse feature of the target is detected by using the ellipse detection method based on arc segment, and an ellipse classification method based on SOR space circle parallelism and perpendicularity constraint is proposed to obtain the reasonable elliptic feature. These features are used to estimate the attitude between the spacecraft and the camera. The experimental results show that this method has good ellipse classification effect and attitude estimation accuracy, and the accuracy rate of classification is not less than 97 for the simulation diagram containing 0 ~ 16% salt and pepper noise. In the experiment, the angle error is not more than 1 掳, the depth direction is less than 10 m, the measurement error is not more than 80 mm, and the other direction measurement error is not more than 15 mm.
【作者单位】： 哈尔滨工业大学电气工程及自动化学院;
【基金】：国家自然科学基金(51075095)
【分类号】：TP391.41;V448
【相似文献】
相关期刊论文 前10条
1 苗育红;周江华;;基于矢量观测的最小方差姿态估计算法(英文)[J];宇航学报;2009年05期
2 傅泽宁;邵晓巍;龚德仁;段登平;;最优递归线性姿态估计算法[J];测控技术;2012年06期
3 罗宇阳;龚德仁;陈筠力;段登平;邵晓巍;;基于罗格里斯参数的航天器递归线性姿态估计算法[J];航天控制;2012年04期
4 李建竹;张骏;;水下晃动平台姿态估计精度仿真[J];西安工业大学学报;2008年01期
5 张根宝;杨永;;粒子滤波器及其在卫星姿态估计中的应用[J];计算机工程与科学;2011年06期
6 郭作敏;飞行器实时姿态估计系统[J];光机电信息;1997年12期
7 李兴唐;郭立红;;基于三维模型的飞机识别与姿态估计[J];武汉理工大学学报;2008年08期
8 闫林波;贾维敏;姚敏立;沈晓卫;;基于扩展卡尔曼滤波的动中通低成本姿态估计[J];电光与控制;2013年07期
9 杨保华;何平;王本利;;一种非线性滤波方法用于卫星的姿态估计[J];控制工程;2006年S1期
10 禹鑫q，
本文编号：2408385
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2408385.html
