Physical information neural network-based light field multispectral temperature inversion system and method

By employing a dual-layer U-Net network and physical information embedding method, the measurement accuracy and robustness issues of optical field multispectral radiation thermometry in complex environments are addressed, achieving high-precision temperature field inversion, which is suitable for temperature measurement in high-temperature and complex environments.

CN121303191BActive Publication Date: 2026-07-14SHANGHAI JIAOTONG UNIV
View PDF 4 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SHANGHAI JIAOTONG UNIV
Filing Date
2025-09-24
Publication Date
2026-07-14

AI Technical Summary

Technical Problem

Existing optical field multispectral radiation thermometry technology lacks accuracy in complex surface characteristics and variable environmental scenarios, and the lack of physical constraints in machine learning models leads to poor interpretability and robustness of the output results.

Method used

A two-layer U-Net network architecture is adopted, embedding custom physical information. Through multiple rounds of optimization iteration and training with composite loss functions, combined with physical constraints such as Planck's law, high-precision and spatially resolved inversion of the temperature field is achieved.

Benefits of technology

It achieves high-precision and reliable temperature field inversion, is suitable for high-temperature and complex environments, improves the efficiency of inversion calculation and the physical consistency of results, and adapts to temperature measurement under complex working conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121303191B_ABST
    Figure CN121303191B_ABST
Patent Text Reader

Abstract

The application provides a light field multispectral temperature inversion system and method based on a physical information neural network, first constructs a deep learning network infrastructure, adopts a double-layer U-Net architecture to extract radiation and spatial distribution characteristics of data; then designs a physical information embedding module, modularizes a Planck radiation law, and guides the network to establish a physical correlation between radiation information and temperature; subsequently, light field multispectral radiation data are collected through experiments, the network is trained after data division and preprocessing are completed, and a mapping relationship between radiation characteristics and temperature characteristics is established; finally, the trained network is migrated to actual test data, and high-precision and rapid temperature inversion is realized. The application has both the advantages of deep learning in processing complex data and the constraint of a physical model, not only improves the precision and calculation efficiency of light field multispectral temperature inversion, but also enhances the applicability of the light field multispectral temperature inversion in a high-temperature complex environment.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Multispectral radiation temperature measurement inversion calculation method based on generalized inverse-neural network, computer and storage medium

    CN113776675A

  • Neuromorphic optical computing architecture system and apparatus

    US20240428063A1

  • Multi-spectral temperature measurement method based on optimization thought

    CN116086617A

  • Temperature profile inversion method based on deep learning

    CN117951485A