基于双目视觉的直升机旋翼桨叶共锥度检测法

Pyramid Angle Detection Method of Helicopter Rotor Blade Based on Binocular Vision

摘要:

提出了一种基于双目视觉的直升机旋翼桨叶共锥度检测方法。首先, 采用平面标定算法对双目视觉系统进行标定, 获取摄像机内外参数; 其次, 采集不同桨叶的立体图像对, 并用Bouguet算法对其进行校正, 获取重投影矩阵; 然后, 对立体图像对中的标记点进行检测与匹配, 获取标记点的视差; 最后, 根据双目视觉测量原理计算出桨尖距地面的高度, 实现桨叶共锥度检测。通过对旋翼塔进行多次测量实验, 结果表明该方法具有动态测量、非接触式、操作简单、危险性小和精度高的优点。

Abstract:

A pyramid angle detection method of helicopter rotor blade based on binocular vision was proposed. Firstly, camera internal and external parameters were obtained by 2D plane calibration algorithm on the binocular vision system. Secondly, stereo image pairs of different blade were acquired and Bouguet’s stereo rectification algorithm was used to rectify them for reprojection matrix. Then, the marked points in stereo image pairs were extracted and matched to obtain the parallax. Finally, the heights of blade tips from the ground were calculated according to the principle of binocular vision measurement, and therefore the blade pyramid angle was detected. The results of the multiple measurement tests on rotor tower show that the method has the advantages of dynamic measurement, non-contact detection, easy operation, less risk and high precision.