Abstract Accurate and efficient measurement of fluid velocities in several kind of experiments, has been acquiring high relevance, for several kind of applications. Particle Image Velocimetry is a noninvasive technique which also provides velocity estimates in several points of the measuring region. Regarding the PIV computations, high resolution of the digital input images, may lead to poor performance. There are also many error sources, regarding the PIV setup itself. This paper describes a method aimed both to speeding up the PIV computations, and minimizing the divergences (noise) present in the output velocity fields. The obtained results are promising, obtaining a speedup of up to 55 % , with respect to the traditional cross correlation based method, with a comparable precision.

中文翻译：

---

一种改善 PIV 中性能和测量误差的方法

摘要 在多种实验中准确有效地测量流体速度，对于多种应用具有高度相关性。粒子图像测速法是一种非侵入性技术，它还提供测量区域多个点的速度估计。关于 PIV 计算，数字输入图像的高分辨率可能会导致性能不佳。关于 PIV 设置本身，还有许​​多错误来源。本文描述了一种旨在加速 PIV 计算和最小化输出速度场中存在的发散（噪声）的方法。获得的结果很有希望，相对于传统的基于互相关的方法，获得了高达 55% 的加速，并具有相当的精度。