论文部分内容阅读
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。
A method based on binocular vision for helicopter rotor blade conicity detection is proposed. Firstly, the calibration algorithm of binocular vision is used to calibrate the binocular vision system to acquire the internal and external parameters of the camera. Secondly, stereo pairs of different blades are acquired and corrected by Bouguet algorithm to obtain the reprojection matrix. Then, Of the points to detect and match, to obtain the parallax of the point; Finally, according to the principle of binocular vision to calculate the tip pitch from the ground height, to achieve a total blade cone detection. The rotor tower was measured several times, the results show that the method has the advantages of dynamic measurement, non-contact, simple operation, small danger and high precision.