Multi-layer perceptron neural network based algorithm for simultaneous retrieving temperature and emissivity from hyperspectral FTIR dataset

Cheng, Jiahui; Xiao, Qing; Li, Xiaowen; Liu, Qinhuo; Du, Yan; Nie, Aixiu

doi:10.1109/igarss.2007.4423824

Cited by 6 publications

(5 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unfortunately, the FTIR measurement method merely aims at point observations and cannot obtain regional BBE estimations. Remote sensing data could be used to measure regional narrowband emissivity but such data cannot represent broadband spectral emissivity [33,34]. The three methods mentioned above exclusively use remote sensing and spectral library data and land surface characteristics vary considerably.…”

Section: Introductionmentioning

confidence: 99%

A New Linear Relation for Estimating Surface Broadband Emissivity in Arid Regions Based on FTIR and MODIS Products

Liu

Mamtimin

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

Broadband emissivity is a crucial parameter for calculating the radiation budget, still, it adopts a constant value in land surface models due to a lack of adequate observations. Arid regions have complex underlying surfaces and estimations of the broadband emissivity in such areas suffer from high spatial variation and uncertainty. Here, we propose a novel method for estimating broadband emissivity in the 8–14 µm range based on Fourier-transform infrared spectroscopy (FTIR) observations, moderate resolution imaging spectrometer (MODIS) emissivity, the leaf area index (LAI) and reflectance products. The proposed method exploits FTIR observations, MODIS single-channel emissivity, reflectance and the LAI to fit a linear regression of the broadband emissivity, so the optimal equation includes emissivity, reflectance and the LAI, with an R2 and root-mean-squared error of 0.942 and 0.08. Then we used the proposed method to generate a broadband emissivity map of Northwest of China, the broadband emissivity estimated by the method showed higher variations and finer distribution in arid areas and sparsely vegetated regions compared to data from the global land surface satellite and land model. An analysis of the relationship between the broadband emissivity, land-use type and soil moisture found an existing but not linear relationship, which indicated that the relationship was complicated under the inhomogeneous surface of wetness and vegetation. In conclusion, our results suggest that the proposed method can accurately estimate the broadband emissivity in arid regions. In future research, we will test the data in a land model.

show abstract

Section: Introductionmentioning

confidence: 99%

A New Linear Relation for Estimating Surface Broadband Emissivity in Arid Regions Based on FTIR and MODIS Products

Liu

Mamtimin

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…In recent decades, considerable efforts have been devoted to solving temperature and emissivity separation (TES). Many algorithms have been developed, including the temperature-independent spectral indices method [20], spectral ratio method [21], two-temperature method [22], reference channel method [23], TES method for Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) [25], gray-body method [26], alpha emissivity method [27], day/night method [28], stepwise refining method [29], piecewise linear regression [30], neural network method [31], [32], and spectral smooth method [16], [17], [33]. For comprehensive reviews, please refer to [18] and [19].…”

mentioning

confidence: 99%

A Modified Interactive Spectral Smooth Temperature Emissivity Separation Algorithm for Low-Temperature Surface

Cao

et al. 2020

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

When ground-leaving radiance from lowtemperature surface is similar to downwelling radiance from the sky, the singular values arise in the retrieved land surface emissivity (LSE) at some specific spectral bands. In addition, too many singular values eventually cause temperature/emissivity separation algorithms to fail. To reduce the occurrence of these singular points, we formulate two indices, including the land-atmosphere radiance contrast index (LACI) and neighbor band contrast index (NBCI). LACI characterizes the contrast between surface radiance and sky downwelling radiance. NBCI characterizes the contrast between the radiance at neighboring bands. These two indices are used as filters to select bands which participate in the iterative spectrally smooth calculation. Thus, we modify the iterative spectrally smooth temperature and emissivity separation (ISSTES) algorithm for low-temperature surfaces. Two methods have been used to evaluate the modified algorithm. First, numerical experiments are conducted to evaluate if the modified algorithm can accurately retrieve the "true" LSE from the simulated data. Second, an artificial low-temperature surface cooled by liquid nitrogen is measured to validate the modified algorithm. The results show that the modified algorithm can effectively avoid singular values, and behaves much better than the original algorithm with errors of less than 0.01 in retrieved emissivity when applied to low-temperature regions, while the modified algorithm brings limited improvement in retrieved temperature.

show abstract

“…This is because of their ability to efficiently handle inherently nonlinear problems. Other examples of neural network technique for the retrieval of atmospheric temperatures using IASI instrument and hyperspectral Fourier transform infrared dataset, respectively, are available in [8,9]. Blackwell [10] very recently did an analysis of high-resolution profiling of atmosphere by considering hypothetical 87 channels microwave sounder using neural network.…”

mentioning

confidence: 99%

Neural Network Based Retrieval of Atmospheric Temperature Profile Using AMSU-A Observations

Mathur

Gohil

Basu

2014

International Journal of Atmospheric Sciences

View full text Add to dashboard Cite

The present study describes artificial neural network (ANN) based approach for the retrieval of atmospheric temperature profiles from AMSU-A microwave temperature sounder. The nonlinear relationship between the temperature profiles and satellite brightness temperatures dictates the use of ANN, which is inherently nonlinear in nature. Since latitudinal variation of temperature is dominant one in the Earth’s atmosphere, separate network configurations have been established for different latitudinal belts, namely, tropics, mid-latitudes, and polar regions. Moreover, as surface emissivity in the microwave region of electromagnetic spectrum significantly influences the radiance (or equivalently the brightness temperature) at the satellite altitude, separate algorithms have been developed for land and ocean for training the networks. Temperature profiles from National Center for Environmental Prediction (NCEP) analysis and brightness temperature observations of AMSU-A onboard NOAA-19 for the year 2010 have been used for training of the networks. Further, the algorithm has been tested on the independent dataset comprising several months of 2012 AMSU-A observations. Finally, an error analysis has been performed by comparing retrieved profiles with collocated temperature profiles from NCEP. Errors in the tropical region are found to be less than those in the mid-latitude and polar regions. Also, in each region the errors over ocean are less than the corresponding ones over land.

show abstract

Multi-layer perceptron neural network based algorithm for simultaneous retrieving temperature and emissivity from hyperspectral FTIR dataset

Cited by 6 publications

References 11 publications

A New Linear Relation for Estimating Surface Broadband Emissivity in Arid Regions Based on FTIR and MODIS Products

A New Linear Relation for Estimating Surface Broadband Emissivity in Arid Regions Based on FTIR and MODIS Products

A Modified Interactive Spectral Smooth Temperature Emissivity Separation Algorithm for Low-Temperature Surface

Neural Network Based Retrieval of Atmospheric Temperature Profile Using AMSU-A Observations

Contact Info

Product

Resources

About