一种基于靶标识别定位算法的视觉测量技术

Vision Measurement Technology Based on Target Recognition and Location Algorithm

视觉测量在飞行试验高精度测试中发挥着重要作用。针对常规视觉测量中多标志点识别定位存在误识别、定位精度低等问题,设计了一种新的环形编码对角标志,并提出了一种基于环形编码对角标志识别定位算法的视觉测量技术。重点介绍了基于EDCircles圆检测算法的标志点识别、基于Harries角点检测算法的亚像素级定位和基于坐标变换的编码信息解码等技术的实现。经过实验验证,研究成果编码标志点识别定位算法识别准确、鲁棒性强、定位精度高,在大投影角度下仍能达到92%的识别率,可以满足飞行试验动态视觉测量任务相关需求。

Vision measurement plays an important role in high-precision test of flight test.Aiming at the problems of false recognition and low positioning accuracy in multi marker point recognition and positioning in conventional vision measurement, a new ring coded diagonal marker is designed, and a multi marker point recognition and positioning algorithm based on EDCircles ellipse detection method technology is proposed.The implementation of marker point recognition based on EDCircles, sub-pixel positioning based on Harries corner detection algorithm, and the decoding of encoded information by coordinate transformation are introduced.Through the experimental verification of the research results, the coded mark recognition algorithm of the research results has the advantages of accurate recognition, strong robustness and high positioning accuracy.It can still achieve 92% recognition rate under large projection angle, and can meet the relevant requirements of dynamic vision measurement tasks in flight test.