相関法と時空間微分法を用いた高解像度画像計測法

概要

論文の詳細を見る
In order to improve the resolution and accuracy, a number of iterative methods have benn proposed, in which re-correlation using the result of initial correlation with discrete window offset is applied. However, in the flow area with large velocity gradient, such as shear flows, compressible flows and turbulent flows, these methods are not very efficient. Furthermore in these methods, the three-point Gaussian peak fit based on correlation method is usually used for a sub-pixel analysis. It is reported that the root-mean-square error of the techniques is of the order of only 0.1 or more pixels and the techniques require very high image density. In this paper, a new high-resolution PIV technique based on cross-correlation and gradient method is proposed. For pixel unit displacement iterative method is applied and then for sub-pixel displacement spatio-temporal derivative method is taken instead of the three-point Gaussian peak fit. The experimental parameters for cross-correlation can be optimized through generalizing the method as a dynamical system, and a sub-pixel displacement is evaluated using spatio-temporal derivative method. The present method is applied for the jet flow images.
社団法人日本船舶海洋工学会の論文
2000-09-25