摘要
为了解决风电叶片双轴疲劳加载中叶片振动状态的检测的实际工程问题,设计了基于单目视觉的精准跟踪方案。通过在视频中选定包含特征标志圆的具有一定特征的ROI区域,设计了基于Mean Shift的图形跟踪算法跟踪视频的R O I区域。并在R O I区域中,应用HO UGH算法,对标志圆进行识别,并识别计算圆心位置。通过这样的算法设计,有效避免了跟踪算法在精确定位应用的不足以及HOUGH算法的误识别问题。在基于模拟视频的实验中,算法计算的特征位置与实际位置之间完全一致,并且满足实时性要求。
In order to solve the actual engineering problem of the vibration state of blades in the wind turbineblades fatigue test,an accurate tracking scheme based on monocular vision is designed in this paper. By choosinga ROI area contains the logo circle selected in the first frame this paper designed a tracking algorithm based onMeanShift to track ROI. In ROI,by applying HOUGH algorithm , the symbol circle can be identified and the circle center position can be calculated as well. By designing such algorithm,it is effective to avoid the fault of thetracking algorithm in the precise positioning of the application and the misidentification of HOUGH algorithm. Inthe experiment based on simulation video,the computation of the algorithm is completely consistent with the actualposition,and the real-time requirement is satisfied.
出处
《装备制造技术》
2017年第11期93-96,共4页
Equipment Manufacturing Technology
