Specific component detection method, determination method, and apparatuses using these methods
AssigneeThe University of Tokyo
InventorsYohei KOBAYASHI, Shuntaro TANI, Hiroharu TAMARU
Life of this patent
Estimating the term…
Abstract
A configuration of the present disclosure specifies a signal point of a specific component and a background point, based on time series two-dimensional spectroscopic image data of a gas; calculates a correlation coefficient between a time series detection signal at the signal point and a time series detection signal at the background point; and generates a time series detection signal of the specific component, based on a corrected time series detection signal obtained by subtracting a product of the correlation coefficient and the time series detection signal at the background point from the time series detection signal at the signal point.
Claims12
A specific component detection method of properly detecting a time series detection signal of a specific component included in a gas, the specific component detection method comprising:
- an input step of receiving an input of time series two-dimensional spectroscopic image data of the gas;
- a specification step of specifying a signal point of the specific component and a background point, based on the time series two-dimensional image data;
- a correlation coefficient calculation step of calculating a correlation coefficient between a time series detection signal at the signal point and a time series detection signal at the background point;
- and
- a signal generation step of generating a time series detection signal of the specific component, based on a corrected time series detection signal obtained by subtracting a product of the correlation coefficient and the time series detection signal at the background point from the time series detection signal at the signal point.
A determination method of determining a state based on a component included in a gas, the determination method comprising:
- applying a learned model, which is obtained by machine learning using a plurality of two-dimensional spectroscopic image data of the gas as an input data set and a state based on the component in the gas as teacher data, to two-dimensional spectroscopic image data of the gas, so as to estimate a state.
Dependencies are inferred from the claim text.