AI 中文总结
研究利用量子计算机在粒子图像测速中的潜力，提出基于多维量子傅里叶变换的QuPIV算法，通过收缩基态投影器增强幅度放大，减少量子电路门数量，经数值研究验证算法端到端能力。
AI 中文摘要
粒子图像测速（PIV）是测量和可视化层流与湍流速度场的主要图像处理技术。通过快速傅里叶变换（FFT）分析互相关，以亚像素精度获取速度场向量。本文提出一种基于多维量子傅里叶变换的量子算法——量子粒子图像测速（QuPIV），取代经典算法计算多达数百万个速度向量。端到端量子算法包括新颖的态制备、改进的幅度放大和输出提取。通过收缩基态投影器增强幅度放大，减少量子电路中的门数量。通过对算法各阶段的数值研究验证了端到端能力。
英文摘要
Particle Image Velocimetry (PIV) is the prime image-processing technique to measure and visualize velocity fields of laminar and turbulent flows. The velocity field vectors are obtained with sub-pixelaccuracy by analyzing cross-correlations, empowered by Fast Fourier Transforms (FFT). Here, we present a quantum algorithm with multidimensional quantum Fourier Transforms, termed Quantum-based PIV (QuPIV), to replace the classical computation of up to millions of velocity vectors. Our end-to-end quantum algorithm includes a novel state preparation, modified amplitude amplification, and the output extraction. We enhance amplitude amplification by a contracted ground-state projector, which allows a significant reduction of the number of gates in the quantum circuit. We justify the end-to-end capability with numerical studies on all stages of the algorithm on both synthetic and experimental data.
Comments15 pages, 4 figures
