基于回归分析的表面流场粒子匹配算法

针对表面流速图像测试中高浓度粒子条件下测量误差大的问题,改进了最邻近匹配算法,提出基于回归分析的表面流场粒子匹配算法,即多帧回归粒子匹配算法。该算法综合粒子距离、运动轨迹、运动趋势等因素,应用一元线性回归分析法拟合粒子中心坐标,通过相关系数完成粒子匹配,实现最佳同源粒子追踪。运用多帧回归粒子匹配算法开发了改进型大尺度模型表面流场图像测试系统(ILSPIV),并将ILSPIV应用到八卦洲河工模型,ILSPIV和旋桨流速仪测量的断面垂线平均流速分布形态及数值基本一致,表明多帧回归粒子匹配算法能提升高浓度粒子条件下流速测量精度,更好地满足工程实践研究需求。

Study on particle matching algorithm for surface flow field based on regression analysis

In order to solve the problem of high measurement error in the condition of high concentration particles in surface velocity image measurement, the nearest neighbor matching algorithm is improved. The particle matching algorithm of multi-frame regression analysis, a particle matching algorithm for surface flow field based on regression analysis is presented. This algorithm takes the particle distance, the motion-trajectory and the movement trends into account. A linear regression analysis is applied to the coordinates of the particle center and the particle matching is accomplished through calculating the correlation coefficient to achieve the best homologous particle tracking. An improved large-scale particle image velocimetry(ILSPIV)was developed with multi-frame regression particle matching algorithm, which had been applied to the Baguazhou Island physical model. The measurement results of the distribution shape and value of for the average vertical velocity between ILSPIV and propeller-type current meter were basically the same, indicating that the particle matching algorithm of multi-frame regression analysis can improve the precision of surface velocity measurement in the condition of high concentration particles and can better meet the needs of engineering practice research.