摘要
基于双目立体视觉标定技术,提供了一种测量旋翼共锥度的新方法。针对桨叶图像需要实时处理的特点,提出一种新的基于Gabor小波变换结合RANSAC方法的特征点匹配算法。首先利用Gabor小波的尺度变换特性提取出桨叶的特征点,然后应用归一化互相关(NCC)对特征点进行初次匹配,最后利用RANSAC算法对初次匹配特征点进行二次匹配,得到匹配点的最优值。以桨叶上边缘点为目标点进行仿真实验,将获得到的桨叶左右图像目标点图像坐标值作为已知解,应用双目立体视觉标定技术求解出三片桨叶顶点的实际三维空间高度值。最后验证了系统的精度。
Provides a new method for helicopter blades pyramid angle measurement by using stereo vision calibration technology.According to the characteristics of which the blades images needs real time processing,this paper presents a new feature point matching method based on Gabor wavelet and RANSAC algorithm.Firstly,using Gabor wavelet scale transform characteristics to pick up feature points,then,first feature points matching take NCC,finally re-matching feature points by using RANSAC algorithm and obtain optimal value.Take the blades top edge as targets and the simulation were carried,use the left and right target ＇s image coordinates value as known value we get the three blades actual 3D height differences by stereo vision calibration technology.At last,the measurement system accuracy was tested.
出处
《微计算机信息》
2011年第1期221-223,共3页
Control & Automation
基金
基金申请人:蔡成涛
项目名称:基于全景视觉的直升机旋翼共锥度测量技术研究
基金颁发部门:国家博士后基金委员会(20090450950)
中央高校基本科研业务费专项资金资助
