Combination of Neural Network and SBFM Algorithm for Monitoring VOCs Distribution by Open Path FTIR Spectrometry

Ren, Yibo; Li, Yan; Yu, Baihua; Wang, Junde; Hu, Lan-Ping

doi:10.1080/10739140601000400

Cited by 6 publications

(3 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, for the limitation of the method, it is hard to find the emission model that is exactly the same to an actual emission. According to the previous studies 19 and the standard on pollution gas detection published by the USEPA, 13 a 2-D Gaussian model was commonly used in gas detection, and the parameters in the model are fitted by PICs and optimization methods. The 2-D Gaussian model used in this paper was obtained according to Eq.…”

Section: Smooth Basis Function Minimization Algorithmmentioning

confidence: 99%

Mapping of methane spatial distribution around biogas plant with an open-path tunable diode absorption spectroscopy scanning system

et al. 2013

View full text Add to dashboard Cite

An open-path tunable diode laser absorption spectroscopy (OP-TDLAS) detector was applied to detect the methane emission from a biogas plant in a dairy farm. Two OP-TDLAS scanning systems were built according to maximum likelihood with expectation minimization (MLEM) and smooth basis function minimization (SBFM) algorithms to reconstruct the two-dimensional (2-D) distribution maps. Six reconstruction maps with the resolution of 30 × 80 were obtained by the MLEM algorithm with "grid translation method" and three reconstruction maps were obtained by the SBFM algorithm with 2-D Gaussian model. The maximum mixing ratio in the first result was between 0.85 and 1.30 ppm, while it was between 1.14 and 1.30 ppm in the second result. The average mixing ratio in the first result was between 0.54 and 0.49 ppm, and between 0.56 and 0.65 ppm in the second result. The reconstruction results validated that the two algorithms could effectively reflect the methane mixing ratio distribution within the target area. However, with the more simple optical rays and less equipment requirements, the OP-TDLAS scanning system based on SBFM algorithm provides a useful monitoring tool of methane emissions in agricultural production.

show abstract

Section: Smooth Basis Function Minimization Algorithmmentioning

confidence: 99%

Mapping of methane spatial distribution around biogas plant with an open-path tunable diode absorption spectroscopy scanning system

et al. 2013

View full text Add to dashboard Cite

show abstract

“…While the paradigm of machine learning and more specifically that of deep neural networks has been applied to spectroscopy-related problems in the past [4] [5], it is to the best knowledge of the authors that multi-gas quantification on passive FTIR spectra has not been attempted before. Ren et al [6] shows rudimentary fullyconnected artificial neural networks as a proof-of-concept on quantifying the path-integrated concentration of captured open-path FTIR spectra. They however, seemingly limited by the technology at their time, model for only 3 gas species and 9 hand-picked data points for the input spectrum.…”

Section: Introductionmentioning

confidence: 99%

Path-integrated concentration and multi-gas detection in FTIR spectroscopy with deep learning methods

Choo

Park²,

Burgstaller

2023

AI and Optical Data Sciences IV

View full text Add to dashboard Cite

The accurate and efficient detection of molecular absorption signatures in FTIR output spectra is a challenging task for traditional filter and statistics-based methods; especially with the quantification of density and robustness to the presence of multiple molecules is concerned. Cross correlation, matched filter and support vector machine techniques generalise poorly to unseen variations of the input. In this work, we employ the powerful embedding capabilities of deep learning models to extract path-integrated concentrations of target gases from the complex spectra generated by HITRAN simulation in the mid-infrared spectrum. A quantitative study is done comparing the applicability of the common neural network types MLP, CNN, and LSTM. The results confirm that convolutional layers are substantially effective at capturing the "spatial" information present in characteristic absorption spectra. Furthermore, we show that such neural networks are robust to noise, temperature and concentration variations, and interference from the presence of other molecules.

show abstract

“…2000,Huang Wengao put forward the new measurement principle of parameters measure by measuring changes in chaotic orbits under the initial conditions fixed [2]. 2006, Litovski Vanco used tent map to achieve A/D conversion, and proposed an experimental circuit [3].2007 ,Ren Haipeng studied chaotic A/D converter based on Bernoulli maps and CPLD [4]. These results had a great boost on the application and research of chaos theory.…”

Section: Introductionmentioning

confidence: 99%

Research on Weak Signal Amplification Principle Based on Parabola Map

Huang

Zhao

Wang

2010

AMR

View full text Add to dashboard Cite

In this paper, a new nonlinear amplification principle based on chaotic theory is proposed. Firstly, according to the basic properties of chaotic systems, such as the sensitivity of the initial conditions and the one-to-one correspondence between chaotic orbit and the initial value, we established the nonlinear enlarge model based on the parabola map. Then, after we studied the nonlinear amplification of the parabola map, and the binary relationship between input and output, we achieved the simulation of the nonlinear amplification with common signals. Thirdly, we compared the result of linear amplification with nonlinear amplification; and discussed the advantages and disadvantages of nonlinear amplification under real situation. Finally, we get the conclusion that the weak signal amplification principle which is based on parabola map has superiority. It can reduce the noise while enlarging the useful signal. In other words, it can improve the signal to noise ratio.

show abstract

Combination of Neural Network and SBFM Algorithm for Monitoring VOCs Distribution by Open Path FTIR Spectrometry

Cited by 6 publications

References 21 publications

Mapping of methane spatial distribution around biogas plant with an open-path tunable diode absorption spectroscopy scanning system

Mapping of methane spatial distribution around biogas plant with an open-path tunable diode absorption spectroscopy scanning system

Path-integrated concentration and multi-gas detection in FTIR spectroscopy with deep learning methods

Research on Weak Signal Amplification Principle Based on Parabola Map

Contact Info

Product

Resources

About