The surface reflectance product is the most crucial and fundamental quantitative product in optical remote sensing, serving as the source for various land surface parameter products. This study initially conducts the retrieval of aerosol optical depth (AOD) from GF1/WFV data. The 6SV model, combined with dark object method and histogram matching, is utilized to achieve AOD inversion at a 16-meter resolution on a per-pixel basis. Subsequently, atmospheric correction is performed using the radiative transfer equation to obtain surface reflectance. The culmination of our efforts involved a comprehensive quantification of uncertainties throughout the production process of the proposed surface reflectance product, allowing for a robust assessment of its reliability, accuracy, and applicability. A thorough evaluation was conducted to ascertain the efficacy and potential of the proposed methodology.

With tens of thousands of channels, infrared ultra-spectral data are expected to improve the accuracy of trace gas profiles, especially carbon monoxide (CO), which possesses a weak adsorption intensity and strong interference signals at the adsorption band. The greatly increased number of channels will generate a considerable amount of redundant information. It is necessary to develop a new method to exclude redundant information, thereby enhancing the retrieval accuracy of ultra-spectral data. In this paper, regarding the ill-posed nature induced by these interference and strong correlation of multi-factors, an improved statistical regression retrieval method is proposed. The ratios of the retrieved gas signal to the interfering signals and the noise equivalent temperature differences (NEΔT) are first analyzed as the weighting factors for the selected channels. By applying the weighting factors in the retrieval process, the proposed method amplifies the contribution of the channels whose information content are more efficient. The proposed method is assessed by the application on the retrieval CO profiles from the simulated ultra-spectral data. The result shows that the root mean square errors (RMSE) of the proposed method for CO profiles is smaller than traditional statistical regression method, and the accuracy is improved by 2.85%.

Land Surface Temperature (LST) is one of the important parameters in thermal environment monitoring. Satellite thermal remote sensing is the major way to obtain spatial-temporal information of LST. However, limited by the cloud contamination and the trade-off between spatial and temporal resolution, current temperature products are difficult to provide all-weather LST. In this paper, a method is proposed to obtain all-weather hourly LST. It consists of two main steps: 1) reconstruction of LST under cloudy-sky by using enhanced annual temperature cycle (ATCE) model and 2) establishment of relationship between LST and air temperature which is used for the acquisition of hourly LST. In the end, the performance of the method is analyzed through the artificial data which is created by masking the origin images. And the results show that the proposed method is valuable for generating all-weather hourly LST.

Directional brightness temperature (DBT) acquired by remote sensing instruments plays a significant role in characterizing the directional anisotropy of land surface, especially for row canopies. The difference between shaded vegetation and sunlit vegetation is ignored in the existing models. In this article, a four-component parameterized directional thermal radiance model (FCPMod) has been proposed to describe the DBT of the row canopy by considering the four components including the sunlit/shaded soil and sunlit/shaded leaf, the improved multiple scattering within the canopy, and the sensor’s field of view (FOV). First, the sensor’s FOV is divided into many tiny rectangles along the row direction and the probabilities of four components in each tiny rectangle are estimated based on the radiative transfer (RT) theory and the bidirectional gap probability. Second, the DBTs are weighted by the four components’ probabilities and brightness temperatures of tiny rectangles. Third, a modified multiple scattering model is proposed to improve the modeling accuracy by considering the contribution of the multiple scattering radiance between soil and canopy. The sensitivity analysis results show that the proposed method performed well compared to the FRA97 model proposed by Françoiset al.(1997) over continuous canopy and the RT model (FovMod) proposed by Renet al.(2013) over row canopy. Finally, the field validations on a maize row canopy show that the proposed FCPMod performed better than about 0.4 K compared with the FovMod.

The advanced hyperspectral imager (AHSI) is one of the sensors aboard the Chinese Gaofen-5 (GF-5) satellite, possessing characteristics of high spatial and spectral resolution, as well as width swath. To better understand the radiometric performance of GF-5/AHIS after its launch, this article presents an on-orbit radiometric calibration approach for AHSI visible and near-infrared (VNIR) and shortwave infrared (SWIR) sensors from field automatic observations with a field spectrometer in the absence of SWIR measurements. A spectrum extension method was proposed to extend the retrieved surface hyperspectral reflectance in the VNIR spectral ranges to SWIR by incorporating the historical hyperspectral reflectance library. The radiometric calibration coefficients of GF-5/AHSI were calculated by linear fitting of the observed digital number (DN) values with GF-5/AHSI and predicted at-sensor radiances with MODTRAN 5 based on extended hyperspectral surface reflectance. Comparisons with onboard calibration results were also performed, and the averaged relative differences were within 5% with$1\delta $standard deviations less than 10% for most bands, except for those in the atmospheric absorption and low signal-to-noise ratio bands. The comparison results indicate that the on-site radiometric calibration results are consistent with the onboard results, and the operational on-orbit radiometric calibration approach is reliable in the case that there are no measurements in the SWIR spectra range. The on-orbit radiometric performance of GF-5/AHSI rapidly degraded during the first several months after its launch and then tended to be relatively stable.

When using a field calibration site to perform in-orbit radiometric calibration of a space-borne remote sensor, the measured signal by the sensor may contain radiation from adjacent pixels, due to scattering in the Earth’s atmosphere and the sensor viewing characteristics. If this is not accounted for in the modeling, the accuracy of the radiometric calibration will be reduced. Nonuniformities of the ground target are more significant for the calibration of high-resolution sensors. In addition, if the brightness contrast of neighboring targets is not small, the impact of the adjacency effects will be more significant. It is important to quantitatively analyze and estimate the influence of this kind of adjacency effects, to reduce the uncertainty of in-orbit radiometric calibration. To evaluate the adjacency effects caused by atmospheric multiple scattering, this article constructed a local atmospheric point spread function model using long time-series satellite-ground synchronous observation data and developed an adjacency effects simulation method, which considers background reflectance spectral information. Tests on Sentinel-2A and Worldview-3 imagery overpassing the Baotou calibration and validation site (Baotou C&V site) (China) indicate that the proposed modeling method can effectively account for the influence of the adjacency effects in vicarious calibration. Uncertainties of relevant parameters and their contributions to the calibration result were also analyzed, and uncertainty assessment results show that the vicarious radiometric calibration scheme considering adjacency effects correction can bring about a total uncertainty less than 7%.

This paper presents a series of temporal vicarious radiometric calibration results of ZY3-MUX sensor over 2016–2019 using the reflectance-based approach. The synchronous ground measuring data have been collected from Baotou sandy site in China given that its high-frequency standard product including surface reflectance and atmospheric parameters. The results show the average difference of radiometric calibration coefficients between calculated results and the official coefficients in each year of ZY3-MUX sensor are 4.6%, 4.6%, 0.64%, 6.28%, respectively. The long-term stability of the radiometric calibration using the data of four years shows a good consistent, and the average difference is 0.92%, 0.64%, 0.50% and 0.65% for ZY3-MUX sensor, respectively. In addition, uncertainty analysis shows that the overall uncertainty for ZY3-MUX radiometric calibration is 4.42%, 4.44%, 4.66% and 3.92%, which also confirms the credibility for radiation quality of Baotou sandy site.

This paper describes an SNO (Simultaneous Nadir Overpass) method based on observation elements matching for radiometric cross-calibration of China HJ-1B and MODIS thermal infrared channels. Firstly, a DTC (Diurnal Temperature Cycle) model with four parameters and ECMWF data are introduced for time matching. Then a BRDF model updated to the TIR domain is built for angle matching. Combining matching coefficients with the spectral matching factor, the TOA radiance of MODIS B31 can be converted into TOA radiance of HJ-1B. Finally, the radiometric calibration results using the images of Qinghai Lake show that the proposed method is effective. Compared with the strict SNO method, the accuracy is improved by 0.73K.

The Chinese Gaofen-1(GF-1) high resolution satellite loaded with four Wide Field of View (WFV) cameras provides observations with high temporal and spatial resolutions. However, the radiometric calibration accuracy of the GF1/WFV should be given when being used to monitoring the earth. In this study, radiometric calibration of the GF1/WFV was carried out first with automatic instrumented Baotou site. Then, the determined radiometric calibration coefficients were cross validated with the MultiSpectral Imager (MSI) onboard the Sentinel-2 satellite. The preliminary results show that the radiometric performances of the four WFV cameras are relatively stable with averaged relative difference less than -1.85%. The standard deviation of the radiometric calibration coefficients during the period from April 2019 to June 2020 is 0.0056, 0.0081, 0.0072, and 0.0076 with respect to the blue, green, red, and near infrared channel. The results of cross validation with Sentinel-2/MSI suggest that the averaged relative difference is -3.16%, -4.28%, -1.15%, and -3.22% with respect to the blue, green, red, and near infrared channel. The results of cross-validation demonstrate that radiometric calibration of the GF-1/WFV cameras using automatic instrumented Baotou site is feasible and operational. And, it is also essential and necessary to update the on-orbit radiometric calibration coefficients of GF-1/WFV cameras during its' entire lifetime for further quantitative application.

The Radiometric Calibration Network (RadCalNet) provide SI-traceable Top-of-Atmosphere (TOA) spectrally-resolved reflectance for automated radiometric calibration of optical satellite sensors. Each RadCalNet site is equipped with automated ground instrumentation in order to provide continuous measurements of both surface reflectance and local environmental/atmospheric conditions needed for the derivation of TOA reflectance values. It is very valuable to use these data in calibrating and motoring optical satellite sensors with the recent launches of very large number of satellites. The work presented here shows the results of vicarious calibration of SuperView-1 satellite sensor using RadCalNet TOA reflectance product. The results also indicated that there was a good consistency among the vicarious calibration coefficients from different RadCalNet sites.

In this study, the ozone column retrieval algorithm is described using the ultraviolet solar radiation from space. The algorithm is based on differential optical absorption spectroscopy (DOAS) technique. Firstly, the differential slant column densities (SCD) of trace gases are retrieved. Secondly, SCDs are subsequently converted to vertical column densities (VCD) by radiative transfer model. Then, the ozone column is retrieved by this algorithm and the sample results show a good correlation with R2of 0.91 and RMSE of 0.89 mg/m2between retrieved and true values. It also can be shown the potential of the algorithm on further atmospheric molecule retrieval in hyperspectral quantitative remote sensing.

Directional reflectance of Baotou sandy calibration site was measured in September 2017. Directional reflectance factors were collected using the Multi-Angles Observation System (MAOS) designed by Academy of Opto-Electronics (AOE), Chinese Academy of Sciences. The directional reflectance was measured at a few viewing and azimuth angles in the 0-30° and 0-360° angular ranges, respectively. Anisotropy and directional effect of surface reflectance were analyzed based on the measured directional reflectance factors. The bidirectional reflectance distribution function (BRDF) model of the calibration site was also modelled, and the model fitting error was approximated to be within 2%. The BRDF model was also used to correct angular difference between ground measurements and satellite measurements in vicarious radiometric calibration carried out in Baotou sandy calibration site, and calibration results with and without angular correction were shown in this work.

Generally, atmospheric correction is a key process before the temperature and emissivity separation (TES). In view of the difficulty and accuracy of acquiring synchronous atmospheric profiles, several in-scene atmospheric correction algorithm have been proposed, one of which is the in-scene atmospheric compensation (ISAC) algorithm. Though this algorithm introduces a good way to find the black-body pixels for enhancing the practicability, it is limited by the spatial resolution of hyper-spectral sensors. This paper tries to apply this method to the thermal band data of Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), since the spatial resolution can preferably satisfy the assumption of homogeneous atmosphere and more homogeneous black-body pixels will be found. The results show that the proposed algorithm is capable of retrieving atmospheric parameters with promising accuracies.

This paper describes a four-component parameterized directional thermal radiance model for row crops, which consists of the thermal radiance of the sunlit/shaded soil and sunlit/shaded leaf, multiple scattering effects of canopy and sensor field of view. The light transmission process of the row crops canopy has been full depicted and the accuracy and sensitivity of the proposed model are discussed in detail. Finally, compared with the FRA97 and FovMod models, the results show that the root mean square error(RMSE) is 0.18K and 0.36K, respectively.

When using field calibration site to perform on-orbit radiometric calibration for a space-borne remote sensor, the observed signal of the sensor may contain energy from adjacent pixels, due to existence of the earth atmosphere and sensor viewing characteristics. Hence the accuracy of radiometric calibration will be decreased to some extent. When calibration of high-resolution sensor is concerned, the non-uniformity of the ground target will be relatively more obvious. In addition, if the brightness contrast of neighboring targets is not small, the impact of adjacency effect will be more outstanding. How to quantitatively analyze and eliminate the influence of this kind of adjacency effect, becomes an actual demand to reduce the uncertainty of on-orbit radiometric calibration. Aiming at the adjacent effect caused by atmospheric multiple scattering, this paper analyzed radiation transfer mechanism first, then constructed a local atmospheric point spread function model using long time-series satellite-ground synchronous observation data, developed an adjacency effect correction method used in on-orbit vicarious calibration, which considers background reflectance spectral information. Test on Sentinel-2A imagery indicates that the proposed correction method can effectively alleviate the influence of adjacency effect in vicarious calibration.

Radiometric characteristics monitoring of optical remote sensing data is a very essential step that enables further quantitative study and application of the data. In this paper, long term vicarious radiometric calibration using ground reflected radiance-based approach was introduced in this study. A case study of Sentinel 2A multispectral imager (MSI) vicarious radiometric characteristics monitoring based on our proposed approach with an automatic observation system was conducted over a large scale desert at the Baotou calibration site in Inner Mongolia, China. And, twelve clear Sentinel 2A MSI scenes as well as ground measurements were successfully acquired during the year of 2017. Top of Atmospheric (TOA) radiance and reflectance were predicted with introduced approach from ground reflected radiance and atmospheric data. The long term radiometric calibration results suggests that the Sentinel 2A MSI display a stable radiometric performance over the calibration period. Vicarious and onboard radiometric calibration results were also cross compared with average relative error about 5%. Uncertainty analysis also show that the TOA radiance overall uncertainty is less than 3.5% due to the atmospheric characteristics, surface characteristics, and the calibration model uncertainties sources.

In this study, the automatic reflectance-based method is used to vicarious radiometrically calibrate the satellite optical sensors using the desert target located in the Baotou site in China. The ground reflected radiance of the desert target were collected automatically using an automatic observation system. The reflectance of the desert target was calculated with the radiance collected with the automatic observation system and the total irradiance simulated from MODTRAN code based on the atmospheric parameters. Then, the TOA radiance can be predicted with MODTRAN code based on the calculated desert reflectance. The automatic reflectance-based approach was applied to the Landsat 8/OLI sensors, and the TOA radiances calibrated by our method were also compared with the observed TOA radiance calibrated with on-board calibrator. Preliminary results show a good consistent and the mean relative difference of the multispectral channels is less than 5%. Uncertainty analysis also show that the TOA radiance overall uncertainty is less than 4% due to the source including the atmospheric characteristics, surface characteristics, and the selected calibration model.

This paper addressed the retrieval of land surface temperature (LST) from combined mid-infrared and thermal infrared data of the Visible Infrared Imaging Radiometer Suite (VIIRS) onboard the Suomi National Polar-Orbiting Partnership (S-NPP). To efficiently remove the effect of the direct solar radiance, a relationship between direct solar radiance and water vapor content, view zenith angle and solar zenith angle is proposed to improve the retrieve accuracy. Then, a split-window algorithm from combined mid-infrared and thermal infrared data is used to correct for the atmospheric effects and retrieve the LST with the aid of emissivity provide by VIIRS product. Finally, comparison of the standard VIIRS LST product and the retrieved LST from the proposed algorithm, a good agreement was shown. Analysis indicated the root mean square error (RMSE) of the LST over these land cover types is 2.04K for desert and 1.84K for vegetation, respectively.

As a novel imaging theoretical, compressive sensing (CS) hyperspectral imaging utilizes the sparse property of the earth objects to efficiently obtain the hyperspectral cube with much less data volume. The construction of sparse basis is of great importance for CS hyperspectral imaging. In this paper, a spectral sparse basis construction method based on earth object's spectral library and redundant dictionary training is proposed. Compared with traditional DCT and wavelet basis, the sparse basis constructed by our method performs much better in simulation experiments.

Land surface temperature and emissivity separation (TES) is a key problem in thermal infrared remote sensing. Current TES methods were proposed and succeeded to apply for the retrieval of land surface temperature and emissivity for the materials with emissivity close to 1. This work addressed the performance of linear spectral emissivity constraint (LSEC) method proposed by wang et al. (2011) for the TES of hyperspectral TIR data for contrast samples (high- and low- emissivity materials). The simulated hyperspectral TIR data are used for analysis and generated with six MODTRAN standard atmospheric profiles by hyperspectral atmospheric radiative transfer model (4A/OP). The influence of initial emissivity estimation is considered in this paper. The results show that initial emissivity estimation has a great impact on the performance of LSEC. LSEC method performs a fairly good result when the initial emissivity is close to the true value, and the RMSEs of temperature and emissivity are smaller than 0.5K and 0.01 when initial emissivity is good. However, the performance is worst when the initial emissivity has a great deviation.

In this paper, great efforts are focused on the temperature and emissivity separation (TES) from hyperspectral thermal infrared data. However, instead of proposing new method, the performances of several published TES methods, including iterative spectrally smooth temperature emissivity separation method (ISSTES), automatic retrieval of temperature and emissivity using spectral smoothness method (ARTEMISS), spectral smoothness method (SpSm), downwelling radiance residual index method (DRRI) and linear spectral emissivity constraint method (LSEC) are analyzed under different instrument characteristics, including the shifting of spectral and the broadening of the full-width half-maximum (FWHM), with the simulated data. The results shows that LSEC has the most robust and accurate performance. DRRI also has a good performance, but a channel selection procedure is required before the use of this method. ISSTES, ARTEMISS and SpSm are more sensitive to the instrument characteristics with some larger errors than other two methods.

Retrieving forest parameters, such as forest height, biomass from polarimetric and interferometric SAR (Pol-InSAR) images has been investigated and well demonstrated via airborne experiments for different types of forest. Terrain slope is a factor that always prevents the wide application of SAR technique due to its slant-looking imaging geometry. It induces changes of backscattering intensity, polarimetric response, etc. Regarding Pol-InSAR forest height retrieval, the effects of terrain slope on retrieval accuracy have been analyzed. The relation between terrain slope and forest height retrieval bias was established through theoretical analysis and simulation procedures based on the PolSARpro software. Based on these results, a simple topographic correction method for forest height retrieval from Pol-InSAR images was presented. This method could alleviate forest retrieval error in rugged areas easily.

This work addressed the estimation of the directional reflectance in Middle Infra-Red (MIR) channel from the data acquired by the Stretched Visible and Infrared Spin Scan Radiometer (SVISSR) onboard Chinese geostationary Meteorological satellite FengYun 2C (FY-2C). SVISSR/FY-2C sensor acquires image covering the whole disk with a temporal resolution of 30 minutes. The MIR directional reflectance retrieval procedure can be seen as follows. Firstly, the atmospheric profiles data provided by European Centre for Medium-Range Weather Forecasts (ECMWF) were used to correct atmospheric influence with the radiative transfer code (MODTRAN 4.0). Secondly, the bi-directional reflectance in SVISSR/FY-2C MIR channel 4 (3.8 micron) was estimated from the combined MIR and TIR channel with day-night SVISSR/FY-2C data. Finally, a BRDF model referred to as the RossThick-LiSparse-R model was used to estimate the directional reflectance in MIR channel from the time-series bi-directional reflectance data. The results have been demonstrated that the method can be applied well to estimate the directional reflectance in MIR channel of SVISSR/FY-2C sensor.

In this paper, an improved method is proposed to simultaneously retrieve land surface temperature (LST), emissivity (LSE) and atmospheric profiles. This method employed the linear spectral emissivity constraint to efficiently reduce the number of retrieved variables. The proposed method was validated with some simulations. The initial guesses were derived from a neural network model. This method could greatly improve the accuracies of LST, LSE and atmospheric profiles. The RMSE of LST was decreased from 5.12 K (the initial guesses) to 1.59 K (the physical retrieved). The retrieved emissivity spectrum was in good agreement with the actual spectrum. An improvement of 1K in the tropospheric temperature was also been found. Those results showed that the proposed method is capable of improving the retrieval accuracies of land surface and atmospheric parameters with the remotely sensed thermal infrared data.

An operational statistical method suitable for nearly real-time estimate of land surface and atmospheric parameters was developed and applied to the Infrared Atmospheric Sounding Interferometer (IASI) observations. The proposed method utilized three steps to solve the ill-posed problems and to stabilize the solution in a fast speed regression manner: 1) the atmospheric profiles and land surface emissivity spectra were expressed by their eigenvectors to reduce the number of unknowns; 2) a ridge regression procedure was introduced to improve the conditioning of the problem and to lessen the influence of noises; 3) a set of optimal channels was selected to decrease the effect of forward model errors or uncertainties of trace gases, and to increase computational efficiency. The retrieval results using the independent simulated data indicate the proposed method is promising. The root mean squared error (RMSE) of land surface temperature is 3.5 K, the RMSE of land surface emissivity at the selected channels is 0.01, and the RMSE of atmospheric temperature and moisture profile are about 2.0K and 0.001g/g, respectively.

MODIS land surface temperature (LST) products have been widely used in numerous applications. Each pixel within the MODIS LST products is acquired at different local solar time even though they are in the same granule. A temporal consistency and spatial comprehensiveness data set will benefit us in the utilization of the LST products in related applications and researches. In this study, a diurnal temperature cycle (DTC) model was employed to normalize the MODIS LSTs to the same local solar time. The MODIS LSTs were derived from the Terra/MODIS and Aqua/MODIS LST products (MOD11_L2 and MYD11_L2, respectively). The results at daytime only are presented because the larger LSTs heterogeneity makes the comparison of LSTs before and after the temporal normalization much clearer. The preliminary results indicate that the spatial variations of the MODIS LSTs caused by different local solar time are removed after the temporal normalization. The temporal normalized LSTs may become more suitable for the analysis of land surface processes.

Total precipitable water (TPW) is an important atmospheric parameter in many applications. A method was proposed to estimate TPW from thermal infrared hyperspectral data. First, 21 channel groups were selected to retrieve TPW. Then, two indices, namely, the differenceand the ratio-depth in each channel group, were used as the measurement of the water vapor absorption. By multivariate regression, the relationship between the TPW and the indices was established. Finally, this relationship was applied to the simulated thermal infrared hyperspectral data. Results showed that the root mean square error (RMSE) of the model is 0.102 g·cm-2, and the relative error is 8.1%. The proposed method needs to be further refined in the future work, including the complete elimination of the Earth's emission in the retrieval.

Owing to the ill-posed problem of radiometric equations, the separation of land surface temperature (LST) and land surface emissivity (LSE) from observed data has always been a troublesome problem. On the basis of the assumption that the LSE spectrum can be described by a piecewise linear function, a new method has been proposed to retrieve LST and LSE from atmospherically corrected hyperspectral thermal infrared data using linear spectral emissivity constraint. Comparisons with the existing methods found in literature show that our proposed method is more noise immune than the existing methods. Even with a NEΔT of 0.5 K, the rmse of LST is observed to be only 0.16 K, and that of LSE is 0.006. In addition, our proposed method is simple and efficient and does not encounter the problem of singular values unlike the existing methods. As for the impact of the atmosphere, the results show that our proposed method performs well with the uncertainty of the atmospheric downwelling radiance but suffers from the inaccuracy of the atmospheric upwelling radiance and atmospheric transmittance, which implies that an accurate atmospheric correction is still needed to convert the radiance measured at the satellite level to the at-ground radiance. To validate the proposed method, a field experiment was conducted, and the results show that 80% of the samples have an accuracy of LST within 1 K and that the mean values of LSE are accurate to 0.01.

This paper makes an attempt to establish a generalized neural network for simultaneously retrieving atmospheric profiles and surface temperature from hyperspectral thermal infrared data. To generate the simulated data covering the whole actual situations, the distributions of surface material, temperature and atmospheric profiles are elaborated carefully. The simulated at-sensor radiances are divided into two sub-ranges, one in atmospheric window and another in water absorption band. The simulated data are transformed in the eigen-domain in both sub-ranges and used as the network inputs. The atmospheric profiles, surface temperature and emissivity are used as the outputs after the eigen-domain transformation. The validation of the trained network indicates that a RMSE of surface temperature around 1.6K, a RMSE of temperature profiles around 2K in troposphere and a RMSE of total water content around 0.3g/cm2can be obtained. The results from the net can be used as initial guess of the physical retrieval model.

Land surface temperature, land surface emissivity and atmospheric profiles are all of great importance in many applications. As the at-sensor radiances are dependent on both the land surface parameters (temperature and emissivity) and atmospheric conditions, it is difficult to simultaneously retrieve these parameters with a high accuracy from multi-spectral radiances measured at satellite level. However, some studies have recently shown that hyperspectral thermal infrared data could be used to derive these parameters simultaneously from space. This paper tries to explore the possibilities to recover with an acceptable accuracy both the geophysical properties and the atmospheric parameters from the hyperspectral thermal infrared data using the neural network technique. The results show that the land surface temperature can be obtained with a RMSE=0.24 K and the atmospheric profiles can also be retrieved with relatively high accuracy. However, further work has to be performed to improve the retrieval accuracy in the near future.