表面流场的PyrLK“角点-质心”法在复杂流场中的应用

Application of PyrLK “corner-mass center” method for surface flow field calculation in the complex flow field

粒子匹配方法是准确获取表面流场的关键。本文详细阐述了新近提出的基于光流算法PyrLK"角点-质心"法的基本原理及具体实现过程,采用该方法对弯道水槽中较复杂的表面流场进行计算,并与目前较为先进的匹配几率法进行了对比。结果表明,对于本文所示的复杂流场,匹配几率法计算结果很大程度上依赖于对运动矢量允许偏差等参数的选取,且还会出现由于提取到非示踪粒子的质心并使其参与计算而产生的误匹配;而PyrLK"角点-质心"法只与运动图像特征信息相关,无示踪粒子特定匹配规则的假定,在复杂流场中能较好地实现粒子的跟踪与匹配。与匹配几率法相比,PyrLK"角点-质心"法对于复杂表面流场计算具有更好的适应性。

The method of particles matching is the key to obtain surface flow field value accurately.This paper expounds in detail on principle and realization course of the newly proposed PyrLK ＂corner-mass center＂ method based on the optical flow algorithm.Using the PyrLK ＂corner-mass center＂ method,the complex surface flow field in the bend channel is calculated,and compared with the advanced method at present called matching probability method.The results indicate that,for the complex flow field shown in this paper,the calculation results of the matching probability method rely on the parameters selected to a large extent such as motion vector permissible variations,and generate the error matching which occurs in implementation process of previous PTV technology owing to mass centers of some non-tracer particles are extracted and used to calculation,because there will be extracted to the center of mass and make them participate in the calculation of tracer particles and the error matching.While the PyrLK corner-mass center method only related to image information features and does not rely on the assumption of tracer particle match rule,it can be better in realizing the tracking and matching of tracer particles in the complex flow.Compared with the matching probability method,this method has a better adaptability in the calculation of complex flow field where the tracer particles have large differences in the displacement vector.