合肥工业大学校徽 合肥工业大学学报自科版

导航菜单

基于深度学习的 NV 色心轴向快速识别

Fast recognition of NV axis orientation based on deep learning

期刊信息

合肥工业大学(自然科学版),2024年11月,第47卷第11期:1481-1485,1492

DOI: 10.3969/j.issn.1003-5060.2024.11.007

作者信息

张文超 $ ^{1,2} $,徐南阳 $ ^{1,2} $,刘玉山 $ ^{1,2} $,周明媞 $ ^{2} $

(1. 合肥工业大学微电子学院,安徽合肥 230601;2. 之江实验室量子传感研究中心,浙江杭州 311000)

摘要和关键词

摘要: 文章提出一种利用 NV 色心的角向偏振光成像图快速确定 NV 色心轴向的深度学习方法。NV 色心的光致发光特性与排列在垂直于 NV 色心对称轴平面上的 2 个正交偶极子有关, 利用角向偏振光可以得到 NV 轴对应的正交光学偶极子的成像图; 通过数字仿真可以模拟不同 NV 轴向在角向偏振光作用下的成像图案, 以此构建数据集并搭建神经网络进行学习; 将神经网络模型作为 NV 成像图和其轴向之间的映射, 从而根据给定的成像图直接得出 NV 色心的部分轴向值, 再依据金刚石晶格中的角度约束关系, 得到金刚石中 NV 色心的完整轴向。基于 NV 色心的角向偏振光成像图, 在保证识别精度的前提下, 实现对 NV 轴向的实时确定, 提高以 NV 轴向为前置条件的各项工作(如磁场重构等)的效率。

关键词: NV 色心;角向偏振光;神经网络;轴向识别;金刚石晶格

Authors

ZHANG Wenchao $ ^{1,2} $, XU Nanyang $ ^{1,2} $, LIU Yushan $ ^{1,2} $, ZHOU Mingti $ ^{2} $

(1. School of Microelectronics, Hefei University of Technology, Hefei 230601, China; 2. Research Center for Quantum Sensing, Zhejiang Lab, Hangzhou 311000, China)

Abstract and Keywords

Abstract: This paper proposes a deep learning enhanced method to quickly determine the orientation of NV axis by using images generated by azimuthally polarized beam. The photoluminescence of NV center is associated with the two orthogonal dipoles which lie in the plane perpendicular to the NV axis. For different NV axis, by activating its dipoles with azimuthally polarized beam, the corresponding image is acquired. Besides, ideal images of NV axis are gotten through numerical simulation, which can be used to build training dataset for neural network. The neural network model is trained as the map between NV image and its orientation to get parts of the orientation information directly from the image. Furthermore, the whole orientation of NV axis in diamond is concluded according to the angle constraint in diamond lattice. By using images generated by azimuthally polarized beam, this paper realizes real-time determination of NV axis orientation with fair accuracy, which elevates the efficiency of works such as magnetic field reconstruction that hold the NV axis orientation as the precondition.

Keywords: NV center; azimuthally polarized beam; neural network; axis orientation recognition; diamond lattice

基金信息

国家自然科学基金资助项目(92265114)

个人中心