摘要
为快速方便地获取直升机旋翼动态三维坐标,提出了一种基于编码标记点和立体视觉的旋翼动态三维测量方法,该方法首先采用平面标定法及Bouguet校正算法实现立体标定和图像校正;其次,依据标记点形状和灰度特征,采用图像二值化和连通域标记等方法检测并定位编码标记点的中心,确定标记点像素坐标;然后,采用邻域扫描法对编码点进行解码,得到标记点的码值,并依据码值实现左右视图中同名标记点匹配;最后,利用投影方程求出标记点三维坐标,实现直升机旋翼动态三维测量。仿真实验结果表明该方法精度高,实际测量误差不超过0.3mm,可满足旋翼动态三维测量的要求。
To obtain 3D coordinate of helicopter rotor rapidly and accurately, a dynamic 313- measurement method based on marked points and stereo vision is presented. Firstly, 2D plane calibration algorithm and Bouguet algorithm are used to calibrate the cameras and rectify the images. Secondly, marked points are extracted from the images by using the methods of image thresholding, connected component labeling and other, so as to acquire the pixel coordinates of the marked points. Then, by using the scanning method, the IDs of marked points are acquired, and then the points within two images are matched according to their IDs. Finally, mapping matrix of the two cameras and the 2D coordinate in the image plane is used to determine 3D coordinates of the helicopter rotor. The simulation results show that the method realizes better precision and the practical measurement error is less than 0.3 mm, which can completely meet the requirements for dynamic 3D-measurement of helicopter rotor.
出处
《半导体光电》
CAS
CSCD
北大核心
2013年第5期904-908,共5页
Semiconductor Optoelectronics
基金
航空科学基金项目(2010ZD56009)
关键词
直升机旋翼
标记点
立体视觉
动态
三维测量
helicopter rotor
marked point
stereo vision
dynamic
3D-measurement
