KEYWORDS: Data modeling, Neural networks, Temperature metrology, Performance modeling, Visual process modeling, Machine learning, Lawrencium, Data acquisition, Coastal modeling, Visualization
Accurate temperature prediction is of great significance to human life and social economy. A series of traditional methods and machine learning methods have been proposed to achieve temperature prediction, but it is still a challenging problem. We propose a temperature prediction model that combines seasonal and trend decomposition using loess (STL) and the bidirectional long short-term memory (Bi-LSTM) network to achieve high-accuracy prediction of the daily average temperature of China cities. The proposed model decomposes the temperature data using STL into trend component, seasonal component, and remainder component. Decomposition components and the original temperature data are input into the two-layer Bi-LSTM to learn the features of the temperature data, and the sum of prediction of three components and the original temperature data prediction result are added using learnable weights as the prediction result. The experimental results show that the average root mean square error and mean absolute error of the proposed model on the testing data are 0.11 and 0.09, respectively, which are lower than 0.35 and 0.27 of STL-LSTM, 2.73 and 2.07 of EMD-LSTM, 0.39 and 0.15 of STL-SVM, achieving a higher precision temperature prediction.
In this study, an improved linear emissivity constraint temperature and emissivity separation (I-LECTES) method was first proposed to overcome the discontinuities problem of the retrieved land surface emissivities (LSEs) in the former linear emissivity constraint temperature and emissivity separation (LECTES) method. Consequently, the hyperspectral thermal infrared data were carefully simulated according to the configuration of Designs & Prototypes microFTIR Model 102, and were used to evaluate the performance of the I-LECTES method. Meanwhile, the I-LECTES method was also compared with the LECTES method. Different the atmosphere and surface circumstances were considered, as well as the different levels of noise equivalent temperature difference (NEΔT). The results showed that the proposed I-LECTES method is of a better accuracy compared with the LECTES method and has the characteristic of keeping the retrieved LSEs continuous, which sounds more reasonable. Because the noises in the ground measured radiance may have more effects on the accuracies of land surface temperature (LST) and LSEs than those in the atmospheric downwelling radiance, the noise in the ground measured radiance should be removed as much as possible to improve the accuracies of retrieved LST and LSEs. Furthermore, taken into account the lower retrieval accuracies for the cold and dry atmosphere, both the I-LECTES method and the LECTES method should be taken a full consideration. The proposed method is regarded to be promising because of its holding continuity and noise-immune.
Land surface emissivity (LSE) is a critical parameter for retrieving land surface temperature (LST) from remotely sensed data. Due to its non-uniformity and a change through vegetation and physical parameters such as texture, composition, surface moisture, roughness, and view angle, the measurement of LSE in laboratory cannot reflect the real world conditions that material interacts with its background and the environment. The filed measurement currently observed by a thermal sensor is radiance, which is a function of many contributing parameters. To accurately obtain the LSE, this paper devotes to develop a scheme for deriving the spectral emissivity from field measured radiance observed by a hand portable FT-IR spectroradiometer Model 102F. A piecewise linear spectral emissivity constraint method is used to decouple the LST and LSE. The results show that the trends of the derived emissivity spectra for different natural surfaces of sand, bare soil and grass are reasonable. Comparisons of several field and laboratory collected LSE spectra for different natural surfaces show that the root mean square errors (RMSEs) are below 0.02, which indicates that the proposed method is accurately to derive LSE spectrum from the measurement of field natural surface with 102F field spectroradiometer.
We evaluate the land surface temperature (LST) generated from the spinning enhanced visible and infrared imager (SEVIRI) onboard the MSG-2 satellite, which was retrieved using the split-window method where the land surface emissivity (LSE) was estimated from the day/night temperature-independent spectral indices-based method. The SEVIRI-derived LST was compared with the MODIS-derived LST extracted from the MOD11B1 V5 product during 7 clear-sky days. The results show that (1) discrepancies exist between the two LST products, with a maximum average difference of 4.9 K; (2) these differences are considered to be time-dependent, since higher discrepancies are observed during the daytime; (3) these differences are land-cover dependent, e.g., bare areas generally present larger differences than vegetated areas; and (4) these differences are inversely proportional to view zenith angle differences. Finally, the main sources of LST differences are investigated and identified in terms of LSE, instrumental noise equivalent temperature difference (NEΔT), and misregistration of the two LST products. The LST differences arising from NEΔT and misregistration are within 0.4 K. Therefore, these discrepancies may mainly result from errors in LSE, which are caused primarily by the atmospheric correction error for the SEVIRI-derived LST.
The goal of this paper is to introduce how to make use of the artificial neural network technique
to develop a new method which can fast recognize atmospheric profiles' characters from hyperspectral
infrared thermal remote sensing. This technique would accelerate the calculation speed of hyperspectral
infrared atmospheric radiative transfer model (RTM). As the launch of hyperspectral infrared sensors such
as Infrared Atmospheric Sounding Interferometer (IASI), it becomes possible for people to take advantage
of the hyperspectral data which contains abundance of precise spectral information, to add constraint
conditions for the researches of some physical models. But in practice, normal hyperspectral infrared
atmospheric RTM are relatively complex and time costing. The calculation speed of these models is not fast
enough to make these models to respond to the variety of atmospheric radiative, or the bright temperature
timely. Therefore, the practical and effective physical models and research methods, such as the practical
surface temperate inversion model, couldn't be founded relay on these transfer models. In order to solve
this problem, institutions and researchers around the world have tried some methods to develop the fast
calculation of atmospheric RTM. But these methods still have problems on speed, accuracy and the
applicability for certain sensors.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.