基于全景视觉技术的旋翼共锥度测量系统的设计及实现  

Design and implement of helicopter blade pyramid angle measurement based on panoramic vision technology

作　　者：姜迈 蔡成涛[1] 邓超[1] 朱齐丹[1] 

机构地区：[1]哈尔滨工程大学自动化学院,哈尔滨150001

出　　处：《吉林大学学报（工学版）》2012年第1期193-198,共6页Journal of Jilin University:Engineering and Technology Edition

基　　金：国家自然科学基金项目(60875025);国防科技工业技术基础科研项目(J192008B001)

摘　　要：在深入分析以往旋翼桨叶共锥度测量手段的基础上,提出了一种基于全景视觉技术的旋翼共锥度测量的新方法。由全景图像球面还原算法原理,推导出了一组用于计算桨叶锥度差的逆投影公式。利用同步电路控制摄像机拍摄实现了对目标桨叶的实时提取,并对全景图像的解算方法进行了相应研究。应用脉冲耦合神经网络(PCNN)对全景图像进行初次分割,重点介绍了在此基础上结合区域标记法对桨叶图像进行特征提取的过程并取得了很好的识别率。实验结果表明,该测量方法对于桨叶锥度差的计算具有较高的精度,可以应用于未来的实际生产过程。According to the investigation of existing rotor blade measuring methods,a new measuring method of blade pyramid angle was proposed based on panoramic vision technology.A group of inverse projection equations were derived from panoramic spherical reductive algorithm to calculate the difference of blade pyramid angles.The target blade real-time extraction was realized by synchronization circuit to control the camera,and the corresponding solving method of the panoramic image was studied.The pulse coupled neural network（PCNN） algorithm was adopted for the initial segmentation of image.Then the feature extraction process of the processed image through region labeling algorithm was introduced to obtain a better recognition rate.Experiment results show that this method has high accuracy and suitable for rotor blade pyramid angle measuring system and can be used in future actual production process.

关 键 词：自动控制技术 旋翼桨叶共锥度 全景视觉技术 逆投影公式 脉冲耦合神经网络 区域标记 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]