基于序列图像的小型无人机姿态快速估计方法与实验  

Fast Attitude Estimation Method and Experiment of Small UAV Based on Image Sequences

作　　者：赵大伟[1] 裴海龙[1] 

机构地区：[1]华南理工大学自主系统与网络控制教育部重点实验室,广州510640

出　　处：《计算机测量与控制》2014年第12期4137-4139,共3页Computer Measurement &Control

基　　金：国家重点基础研究发展计划资助项目(2014CB845301/2/3);2013年度开放基金项目(1214);华南理工大学中央高校基本科研业务费自助项目(2014ZP0021)

摘　　要：为自主、精确地获得无人机的飞行姿态,文章研究并提出了一种利用无人机系统搭载的摄像机获取的序列图像来快速获得小型无人机飞行姿态的方法;该方法首先应用SIFT算法,通过多线程并行处理提取相邻图像特征点,然后依据设计的变换模型由匹配的特征点估计出帧间图像变换参数,并利用RANSAC算法剔除错误点,最终解算出机载平台的姿态角变化量;最后通过比较实验测得的姿态数据与实际姿态数据,证明了该方法的有效性。In order to accurately and independently measure the attitude of small UAV,a fast way to acquire the estimation of the attitude based on the camera equipped on the small UAV is introduced in this paper.At first,SIFT algorithm is used to extract series of the interest points from inter -images,and then based on the matching feature points to estimate the inter image transformation model parameters with Least Squares method.RANSAC algorithm is used to remove the error pairs of matching interest points.Finally,the transformation matrix between the inter-images is es tablished to obtain the attitude angles.By comparing the experimentally measured posture data and the actual posture data,this method was proved to be available.

关 键 词：SIFT RANSAC 特征匹配 序列图像 姿态估计 

分 类 号：TP319.4[自动化与计算机技术—计算机软件与理论]