.png){kind=icon}
.png){kind=icon}
Return
A method for processing multispectral radiometric thermometry data based on BP-Alpha constraints
.png){kind=icon}
PRE
AI
DOI：10.1016/j.infrared.2024.105387.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Multispectral radiometric thermometry is a commonly used method for infrared temperature measurement. The biggest challenge in processing multispectral data is the unknown emissivity. The traditional method of modeling emissivity assumptions is inadequate in dealing with emissivity instability, particularly when there is a rapid change in emissivity due to varying degrees of oxidation during metal heating. To address the temperature measurement challenge, this paper converts the multispectral temperature measurement problem into a constrained optimization problem. This is achieved by constructing a new type of constraints, using a BP neural network to establish the emissivity range constraints, and constructing an Alpha spectral model to establish the emissivity shape constraints. The improved Lichtenberg algorithm is then used to solve the problem, resulting in significantly improved temperature inversion accuracy. The method's effectiveness was verified through simulations and GH3044 alloy experiments. The temperature inversion's maximum absolute and relative errors were 9.5 K and 0.92%, respectively.
Keywords:
Emissivity
Multispectral thermometry
Constraint optimization
Journal
I
IF:
3.4
Papers:
5.8K
Citations:
1.2W