一种基于HRUnet网络的高精度相位解调方法

By directly demodulating the high-precision unwrapped phase distribution from the cosine interferogram using the HRUnet network, the problems of complex devices and error correction in existing technologies are solved, and high-precision phase measurement of dynamic samples is realized.

CN116428965BActive Publication Date: 2026-07-17XIAN TECH UNIV

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
XIAN TECH UNIV
Filing Date
2023-03-24
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

In the existing technology, phase demodulation methods have problems such as complex equipment, complex error correction, and inability to be used for measuring dynamically changing samples. In particular, there are obvious abrupt phase errors in the unwrapped phase distribution directly output by the network.

Method used

Phase demodulation is performed using the HRUnet network. The HRUnet network is constructed and trained by generating data using Zernike polynomials. Using cosine interferograms as input, it directly outputs a high-precision unwrapped phase distribution, avoiding a complex error correction process.

Benefits of technology

It achieves high-precision unwrapped phase demodulation, is suitable for dynamic sample measurement, simplifies the operation process, improves the accuracy and precision of reconstructed phase distribution, and is applicable to closed interferograms with different sparsity levels.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116428965B_ABST
    Figure CN116428965B_ABST
Patent Text Reader

Abstract

本发明涉及光学干涉测量领域,具体公开了一种基于HRUnet网络的高精度相位解调方法。只需一幅干涉图,即可得到无包裹相位信息。本方法直接从单幅含有闭合条纹的干涉图得到无包裹相位分布,先建立并训练神经网络HRUnet,然后把余弦干涉图Icos输入训练好的网络,输出得到无包裹相位分布。具体包括如下步骤:1.数据准备,使用泽尼克(Zernike)多项式产生训练数据和测试数据用于网络训练;2.建立HRUnet网络,用步骤1得到的数据对网络进行训练;3.把余弦干涉图输入训练后的神经网络,输出得到无包裹相位分布。本发明提出的技术一步输出无包裹相位,相位重建精度高,在光学干涉测量研究领域有着很大的应用前景。
Need to check novelty before this filing date? Find Prior Art