Optical structure optimization design method based on deep neural network

A technology of deep neural network and optical structure, applied in the field of optical structure design, can solve the problems of long simulation time of optical structure, low efficiency of manual adjustment, and difficulty in finding global optimal structure parameters, so as to achieve the effect of improving optimization efficiency.

Pending Publication Date: 2021-11-02
WUHAN UNIV OF TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The present invention solves the problems in the prior art that the simulation time of the optical structure is long, the efficiency of manual adjustment is low, and it is difficult to find the global optimal structural parameters by providing an optical structure optimization design method based on a deep neural network.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Optical structure optimization design method based on deep neural network
  • Optical structure optimization design method based on deep neural network
  • Optical structure optimization design method based on deep neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0075] Embodiment 1 provides an optical structure design method for polarization conversion, see Figure 4 , mainly including: acquisition and preprocessing of the simulation data set, construction and initialization of the spectral prediction network corresponding to the polarization conversion efficiency, training of the spectral prediction network, multi-valued adaptive particle swarm optimization algorithm to optimize the optical structure, and simulation verification.

[0076] For the data set preprocessing part, the obtained structure data is average-normalized so that the data size of the structure parameters is (-1,1).

[0077] The spectrum prediction network corresponding to embodiment 1 is as Figure 5 As shown, the spectrum prediction network based on the deep neural network can accept each structural parameter of the optical structure as input data and predict the corresponding transmission spectrum in a very short time. Among all the modules of the spectral predi...

Embodiment 2

[0097] Embodiment 2 provides a main flow of an optical structure design method for optical sensing, see Figure 8 , similar to Example 1. The main difference between embodiment 2 and embodiment 1 lies in the specific number of layers and parameter settings of the spectrum prediction network, and the specific setting of the objective function.

[0098] The main function of the input layer IN is feature extraction, which is similar to the modeling process of simulation software. The number of layers of the input layer IN and the number of units in each layer mainly change according to the complexity of the optical structure. The main function of the intermediate residual network RN is The function is nonlinear fitting, which is similar to the process of solving Maxwell's equations in simulation software. The number of layers of the intermediate residual network RN and the number of units in each layer mainly change according to the complexity of the spectrum to be fitted. The m...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention belongs to the technical field of optical structure design, and discloses an optical structure optimization design method based on a deep neural network, and the method comprises the steps: taking a parameter of a to-be-optimized optical structure as an input, and employing a spectrum prediction network based on the deep neural network to carry out the prediction of a spectrum; and calculating and obtaining a target function based on the predicted spectrum, optimizing the target function by adopting a multi-valued adaptive particle swarm algorithm, and adaptively and reversely searching an optimized optical structure. The problems that in the prior art, simulation time of an optical structure is long, manual adjustment efficiency is low, and global optimal structure parameters are difficult to find are solved, and the optical structure with the optimal optical attribute can be obtained conveniently and efficiently.

Description

technical field [0001] The invention belongs to the technical field of optical structure design, and more specifically relates to an optical structure optimization design method based on a deep neural network. Background technique [0002] In recent years, by manipulating the interaction between light and optical micro-nano structures, the field of optics has undergone revolutionary changes, such as holography, polarization conversion and optical sensing, etc. The performance of devices involving metasurfaces has been improved. How to design an optical structure to optimize the optical properties has attracted much attention. [0003] In optical design, optical properties such as reflection, transmission or absorption spectrum depend on the structural parameters and material refractive index of the unit structure. When the selected material and light source range are determined, the optical properties directly depend on the structural parameters. Therefore, in the tradition...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06F30/20G06N3/00G06N3/04G06N3/08
CPCG06F30/20G06N3/08G06N3/006G06N3/045
Inventor 杜庆国喻佳成陈志伟王原丽李政颖高雅玙
Owner WUHAN UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap