基于密度聚类分析的粒子测速算法

New Method for Particle Velocimetry Measurement Based on Density Clustering Analysis

提出了一种具有密度预测分析的粒子图像测速算法(Density clustering particle image velocimetry,DCPIV)。在对粒子图像进行预处理时,根据图像中粒子分布的复杂性以及流体运动的连续性,首先使用密度聚类法对粒子图像按密度进行划分,然后对不同区域的密度图进行相关处理,并将此结果作为粒子相关或者粒子追踪的预测值进行处理。最后,使用合成的粒子图像进行了方法验证,结果表明:该方法能够解决实际操作时粒子分布不匀的问题以及由于旋转所引起的误匹配问题,并且处理结果具有较高的精度。

A new method, called the DCPIV（Density clustering particle image velocimetry）, for particle image velocimetry based on density clustering analysis （DCA） is proposed. The nearest-neighbor and correlation methods are combined to achieve movement measurement by DCA. Firstly, the particle images are divided into different regions. Secondly, the matching method is used in these regions, and the matching results are used as the estimates for particle correlation velocimetry （PCV） or particle tracking velocimetry （PTV）. Finally, synthetic particle images are tested. Experimental results show that the proposed method is robust to measure particle image movement and the velocities of the field are obtained. The processing resuits have high precision.