Mangroves are important plant communities in coastal ecosystems， and have enormous social， ecological and economic values. The development of remote sensing technology provides an efficient and convenient way for mangrove monitoring. Radar remote sensing has a unique advantage in mangrove distribution area because it has high penetration and is unaffected by cloud and rain. This paper reviews the study on mangrove monitoring based on radar remote sensing in recent decades in the aspects of mangrove scattering mechanism， mangrove classification and recognition， and mangrove biophysical parameters retrieval. The summary and comparison of different methods in three aspects are also proposed. Finally， the future improvements are discussed according to the existing problems.
Soil moisture is considered as an important state variable of terrestrial ecosystems and water cycle， which plays an important role in many researches， such as vegetation growth monitoring and crop yield evaluation. In order to eliminate the effect of vegetation scattering and achieve high-precision retrieval of farmland soil moisture， time series of Sentinel-1 data and MODIS products were used as experimental data to retrieve soil moisture. The advanced integral equation model was coupled with ratio vegetation model， which is parameterized by four different optical vegetation parameters and VH cross-polarization backscattering coefficient for vegetation scattering contribution， to eliminate the impact of vegetation scattering and then achieve high-precision inversion of soil moisture. The results show that the coupled model can simulate the Sentinel-1 VV-polarized backscattering coefficient but the soil moisture retrieval results are not ideal with the maximum correlation coefficient （R） is 0.54 when VH is used for parameterizing the scattering contribution of vegetation. Different from VH， the overall soil moisture retrieval results is relatively better with a maximum R of 0.79 when four optical vegetation parameters are used. However， there is significant spatial difference at different stations with the results of optical vegetation parameters， with R ranging from 0.07 to 0.79. Therefore， in future results， it is better to combine radar data and optical data for eliminating the scattering contribution of vegetation and realizing the high-precision retrieval of soil moisture and observation of dynamic changes in vegetated areas.
Soil moisture is a key variable in quantifying water and heat exchanges between land surface and atmosphere， which also affects the partitioning of surface sensible and latent heat fluxes， and estimations of water budget and vegetation transpiration. Study of soil moisture on the Tibetan Plateau is of great significance to improve the simulation of water and energy budgets on the plateau. After the launch of SMOS and SMAP satellites， L-band passive microwave remote sensing has become the main way of monitoring soil moisture at large scale. This paper reviews and summarizes recent progresses on the L-band microwave radiometry observation and soil moisture retrieval over the Tibetan Plateau， including measurements and simulations of brightness temperature based on ground-， aircraft-based and spaceborne platforms， development of regional-scale soil moisture monitoring networks， evaluation of satellite products and development of soil moisture retrieval algorithms. Based on the reviews， we summarize the main problems exist currently on simulating the L-band microwave emission and retrieving soil moisture on the plateau， such as lack of evaluating microwave emission simulation and improving satellite-based soil moisture retrievals at plateau scale， and absence of soil moisture products for frozen soil conditions. In view of above existing problems， this paper further suggests that future work should pay more attention to improve the L-band microwave emission simulation and soil moisture retrievals at the plateau scale， and to enlarge the applications of soil moisture products， such as to improve the understanding of plateau-scale water and energy budgets， vegetation growth and drought monitoring.
The cosmic-ray is a non-destructive method of measuring soil moisture at the 100-meter scale. Using multi-layer observations from a dense monitoring network， this study explores the retrieval of soil water content through fast neutron time series obtained from COsmic-ray Soil Moisture Observing System （COSMOS） in a typical Karst watershed in Qingmuguan， Chongqing. Some specific treatments/investigations are conducted toward improving the overall retrieving accuracy， including ①using Savitzky-Golay filter to smooth the fast neutron time series； ②analyzing the role of vegetation water content， and ③comparing different data screening schemes during the calibration and verification phrases. Results suggest that the vegetation water content has negligible impacts on the COSMOS retrieving in this specific area， and the calibration by considering longer and different soil moisture records delivers the best agreement with the ground truth. Finally， the calibrated algorithm was applied to the whole COSMOS measuring period to produce a complete soil moisture record， which is further indirectly verified with neighboring soil moisture and precipitation observations， and help reveal the seasonal soil water content variations. In general， the proposed COSMOS soil moisture retrieval algorithm shows robust applicability and is expected to support regional scale of long-term soil moisture monitoring and hydrometeorological studies in this region.
The Huaihe River Basin is selected as the research area. Based on the daily average data of 313 soil moisture observation stations in the Huaihe River basin from June 2016 to May 2019 ， the soil moisture products （L2_SM_P_E） of SMAP （Soil Moisture Active Passive） with 9 km resolution were assessed by using a variety of indicators. In addition， the influence of vegetation， soil， topography on accuracy were discussed. The results show that： （1） Generally， L2_SM_P_E cannot reach the expected accuracy of 0.04 m3/m3 in Huaihe River Basin， which has the characteristic of overestimating in wet area and underestimating in dry area， but it can better reflect the spatial distribution characteristics of soil moisture in the basin， and also can indicate the high wet area and low wet area. （2） There are obvious regional and seasonal differences in L2_SM_P_E accuracy. The accuracy in winter is obviously better than that in other seasons， the unbiased root mean square error （ubRMSE） in most areas of the basin is close to the expected accuracy.In some northern parts of the basin and Funiu Mountains and Dabie Mountains，it has reached the expected accuracy. In spring and autumn， the accuracy of the northern part of the basin and Dabie Mountains is higher. In summer， the availability of L2_SM_P_E is poor. （3） L2_SM_P_E have good consistency with precipitation， and its response to precipitation is more sensitive than the observed value of soil moisture. During and after precipitation， the error of L2_SM_P_E is mainly random error； when the soil is relatively dry， it is mainly negative systematic error. （4）The accuracy of L2_SM_P_E is not closely related to the soil type at the sampling point. The accuracy of mountain areas is better than other areas.
The L-band microwave radiometer is an important experimental equipment for detecting soil moisture and Sea Surface Salinity（SSS）. However， the increasingly Radio-Frequency Interference（RFI） makes the L-band passive microwave remote sensing unable to meet the inversion accuracy requirements of ground surface parameters. In this paper， an integral type L-band microwave radiometer with fast sampling capacity is mounted on the vehicle mobile platform. The ground measurement is carried out in the Shandian River Basin of North China. The Asynchronous Pulse Blanking algorithm （APB） and the median comparison algorithm based on the coefficient of variation are used to detect and mitigate the radio-frequency interference of the measured results. Both of two algorithms have achieved a certain effect on radio-frequency interference detection. According to the test results， the RFI observed by the L-band ground-based microwave radiometer in the remote sensing experimental area of the Shandian River is mainly small-scale pulsed interference， the field interference level is generally 3~4 K， most interference persist 1~2 ms， and the interference rate is between 2% and 14%. The analysis shows that the field interference magnitude in the experimental area is significantly smaller than that in the cities and towns， and the interference to V polarization is slightly more serious than that to H polarization. It is found that the APB algorithm has more false detection phenomena or False Alarm Rate（FAR）， while the median comparison algorithm based on the coefficient of variation can more tolerate the radiation brightness temperature fluctuation of the target itself， and the minimum detection interference level or magnitude of the latter algorithm is less than 3 K. The measurement data of ground-based L-band passive microwave radiometer will contribute to the verification and calibration of remote sensing data of both aviation and satellite.
Research on downscaling Soil Moisture （SM） provides a technical means to solve the coarse resolution problem of passive microwave SM products and to better serve small-scale regional applications. On the basis of SMAP SM products and multiple remote sensing auxiliary data including the MODIS product and so on， supported by the land use data， a hybrid downscaling method was developed by combining the parameter statistics and spatio-temporal fusion downscaling method. And taking Oklahoma， USA as the study area， the downscaled results were evaluated using SMAP 9 km products and in-site data. The results show that： the hybrid downscaling method could obtain downscaled results with rich details and complete spatial coverage. Compared with the two single downscaling method based on parameter statistics or spatio-temporal fusion， the spatial distribution of the hybrid downscaled result was the most similar to the real SMAP 9 km product， and it has the highest temporal accuracy against site data whether on the whole or under different land use. Therefore， the proposed hybrid downscaling method combining parameter statistics and spatio-temporal fusion was feasible.
Solar-Induced chlorophyll Fluorescence （SIF） is an ideal indicator of global vegetation productivity. Although there is still no satellite-based sensor designed for SIF monitoring specifically， there are a series of atmospheric monitoring hyperspectral sensors which have potential for SIF retrieval. And a number of satellite-based global SIF products have been developed and published. Furthermore， some spatial and temporal extended SIF products have also been developed to better match the requirements of SIF application. The design of specific satellite-based SIF sensors is already in progress in both China and Europe. Although the products of satellite-based SIF products developed fast in recent years， lots of uncertainties and limitations remains for application. In this paper， the existing and in-coming satellite-based sensors for SIF detection， the published global SIF products were summarized. From the perspective of application requirements， the existing limitations of global SIF products and the development direction in the future were analyzed. This paper can serve as a reference for the application of existing SIF satellite products and the design of future satellite-based SIF exploring missions.
According to the characteristic that Solar-Induced chlorophyll Fluorescence （SIF） can effectively indicate the water stress of land surface vegetation， we proposed a Normalized Solar-Induced Chlorophyll Fluorescence Drought Index （NSDI） for winter wheat drought monitoring in the Huang-Huai-Hai region. First， the original SIF data retrieved by the Sentinel-5p Tropospheric Instrument （TROPOMI） were processed into spatially continuous data with a spatial resolution of 0.1 degree. Missing values were then filled via the linear interpolation based on time series analysis， and S-G filters were applied to reconstruct high spatial and temporal resolution SIF dataset. The NSDI is developed using this reconstructed SIF dataset and winter wheat distribution data. The analysis of typical drought events revealed that the NSDI and the Normalized Difference Vegetation Index （NDVI） are strongly correlated with the R2 of 0.60， the NSDI and the temperature vegetation drought index （TVDI） are also strongly correlated in different mature regions， with the highest R2 of 0.66 in Yanshan region， and the lowest R2 of 0.44 in Huanghuai plain region. The NSDI index is also highly correlated with the in-situ soil moisture data， with an R2 of 0.53 and 0.54 respectively in Hebei and Shandong sample area， and an overall R2 of 0.51. Analysis of monitoring data from the Internet of Things shows that the NSDI index can respond to changes of drought within a lag period of less than 2 days， and its change trend is highly correlated with soil moisture in the field. The experimental results show that the NSDI index can effectively indicate the drought of winter wheat in Huang-Huai-Hai region from the spatiotemporal perspective.
The lunar target has the characteristics of small atmospheric influence and obvious difference from the background temperature. It can be used as a very stable reference source for band registration to analyze the imaging positions of different bands. Taking the second generation Medium Resolution Spectral Imager （MERSI-Ⅱ） on FY-3D as an example， a method of on-orbit registration using the moon is introduced. By preprocessing the moon image of each band of MERSI-Ⅱ， and adopting the method of sub-pixel maximum correlation iteration for image registration， the registration parameters of sub-pixel accuracy are obtained. The registration results were evaluated and analyzed by comparing the centroid distance of the moon and the difference of gray level. The results show that the lunar registration method has high accuracy and the consistency with the lunar centroid distance is better than 0.05 pixel. The registration based on the moon has the advantages of simple and efficient， and is not affected by the spectral differences of instrument bands. It is stable in the registration of absorption bands， and is significantly better than other methods.The on-orbit registration results of MERSI-Ⅱ based on pre-launch test are compared and verified by the lunar registration results. It is found that the on orbit registration deviation of MERSI-Ⅱ is relatively small as a whole， however， there is 0.5pixel deviation in the scanning direction of the shortwave band， which has an impact on the subsequent high-precision quantitative products. The method can be used to find the long-term change law of band registration， which lays a foundation for the re-calibration of historical data and the subsequent development and application of quantitative remote sensing products.
Precision agriculture and intelligent agriculture need high precision terrain factor data of field scale， but the existing topographic mapping data of arable land can not meet the demand. In order to establish a high spatial resolution Digital Elevation Model （DEM）， obtained SPOT-6 multispectral data in June， July， August and September 2016， SRTM DEM and the actual elevation of the study area were measured. High resolution DEM（6 m） data are obtained by Kriging spatial interpolation based on the measured elevation. Taking the grid value of SRTM DEM and Normalized Difference Vegetation Index （NDVI） as inputs， the multiple linear regression and BP neural network reconstruction model are established， The verification is based on the measured ground elevation and the DEM data obtained by UAV. Compared with the UAV DEM and ZY-3 DEM. The results show that： ①The accuracy of the linear regression model introduced into NDVI time series is 96%. The RMSE is 1.12； The accuracy of BP neural network model is as high as 98.7%， and RMSE is reduced to 0.86. ②The temporal and spatial variation of NDVI in growing season is the result of topographic factors such as slope and slope position. ③The improved SRTM DEM based on BP neural network achieves higher spatial resolution， which is similar to the spatial trend of UAV DEM， and its accuracy is higher than that of ZY-3 DEM. It can provide data support for field variable management， precise management zoning， soil classification and fine mapping.
Earth observations of satellite or airborne sensors are easily interfered by the atmospheric conditions， thereby resulting in the frequent cloud contamination in the acquired images， reducing the availability and validity of data. In this paper， a thin cloud removal method based on intra-class linear regression is proposed for visible remote sensing images， which mainly consists three steps. Firstly， the local dark object （minimum in local） is searched band by band with a certain window size. The dark objects samples are then used to regress the linear correlation of clouds among bands. Secondly， the cloud correlations among bands are combined with the cloudy image model to generate the synthetic image without cloud contamination， and the K-means classification is performed on it to obtain the land cover types. Based on that， the linear relationship of different land covers can be estimated using the corresponding clear samples. Thirdly， by integrating the linear correlations of clouds and various land covers， the clear surface information can finally be solved from the cloudy image model. Both the simulated and real data were collected to validate the effectiveness of the method from visual and quantitative aspects. Experimental results demonstrate that the thin clouds in various scenes can be totally removed by the method and the degraded information can be recovered satisfactory.
In order to evaluate the elimination performance of topographic shadow effect by the Shadow-Eliminated Vegetation Index （SEVI），four multispectral images of the Sentinel S2B （10 m）， GF-1 （16 m）， Landsat 8 OLI （30 m） and GF-4 （50 m） were used to study. The Normalized Difference Vegetation Index （NDVI） calculated from the surface reflectance and sun-canopy-sensor （SCS）+C correction was used to compare with SEVI calculated from the surface reflectance data. The evaluation methods included the vegetation indices value analysis， absolute relative error analysis， Coefficient of Variation （CV） analysis and scatter plots analysis of the cosine of solar incidence angle （
In the data assimilation method， the observation error covariance matrix is correlated and dependent on time and state. in view of such correlation characteristics， the robust filtering method is combined with the estimation of observation error covariance to obtain the time-varying covariance of observation error， and the robust data assimilation method with observation error estimation is proposed to update the observation error covariance and improve the estimation performance. In this work， nonlinear lorenz-96 chaotic system is used to evaluate the robustness and assimilation accuracy of robust filtering with observation error estimation and original robust filtering under three different optimization methods. The performance of the two methods is compared and analyzed when the model error， the number of observations and the performance level coefficient change. The results show that the observation error estimation technique can improve the accuracy of the state estimation， and the robust data assimilation with the observation error estimation is more robust on the change of system parameters.
Monitoring and early warning of sub-healthy forests caused by forest diseases and insect pests and other disturbance types can not be carried out in time， resulting in a passive situation （disaster-relief /post-disaster） for a long time. Based on the multi-temporal GF-1 WFV data from May to September 2019， this paper uses the ratio vegetation index and the red-green vegetation index to monitor "disaster" information such as reverse growth， leaf canopy stress or loss of color in quasi-real time. The results show that although the degradation of chlorophyll such as withering and wilting of tree leaves and gradually transforming into lutein and red leaf pigment requires a certain process， or the "disaster symptoms" sometimes have a lag， but the high frequency remote sensing dynamic monitoring results are useful for guiding the ground inspection of forest disasters. It has a positive effect on improving monitoring coverage and scientificity， and preventing large-scale disasters. The high revisit cycle of domestic GF-1 and GF-6 WFV remote sensing data provides a solid data guarantee for the monthly monitoring of the growth process of forest resources， and meets the needs of hectare-level leaf growth and degradation early warning monitoring.
Monitoring urban expansion and environmental change is significant to promote high-quality integrated development in Yangtze River Delta urban agglomeration （YRD）. Based on DMSP/OLS night-time light data， multi-source remote sensing environmental indicators， methods of slope trend line analysis， construction of Comprehensive Evaluation Index （CEI） were used to analyze the characteristics of urban expansion， environmental change and their coordination of YRD from 2001 to 2013. The results showed that： （1）46.24% of the cities expanded significantly， with the highest level of urbanization and urban integrated development in the region of Shanghai-Suzhou-Wuxi-Changzhou， followed by the Hangzhou Bay urban agglomeration and the regions along the Yangtze River. However， the high-speed urban expansion mode was not multi-regional and large-scale in YRD. GDP， the degree of opening to the outside world and electricity consumption were the main driving factors for the high-speed urban expansion. （2） In 46.35% of the significant urban expansion area， the environment gradually degraded， which was concentrated in Jiangsu and Anhui Province. Environment-degradation type was clustered around the downtown of Hefei. Environment quality of Zhejiang Province was the best， while that of Shanghai was almost unchanged. PM2.5 growth was an important reason for environmental change in most cities. （3） The superposition of urban expansion and environmental change can reflect the coordination degree of urban development and ecological environmental change. The coordination in Jiading District， Pudong New Area and Chongming Island in Shanghai， Suzhou-Wuxi-Changzhou and cities along Yangtze River was weak. The "gradual" distribution pattern in Zhejiang Province cities reflected high coordination， while that in Hefei city reflected a non-coordination mode that the environmental quality was far behind urbanization process. The results of high-resolution satellite images positioning showed that the combination of DMSP/OLS and CEI can successfully monitor the surface cover and its transfer that cause serious environment damage in the process of urbanization. This work can reveal the status of regional sustainable development， and has an instructive role in the future study of spatial transfer of dynamic development coordination between urbanization and environmental quality in the Yangtze River Delta.
Many current studies use remote sensing images to pay attention to the dynamic changes of surface water， but the changes obtained by different algorithms need to be further analyzed. Based on the GEE platform， taking Three River Headwater（TRH） as an example， using JRC global surface aquatic products to analyze the dynamic change of the surface water area of TRH from 2001 to 2018， and combining the atmospheric reanalysis products ERA5 and PERSIANN-CDR for driving force analysis. The results show that during the period 2001-2018， the permanent water area of the TRH increased from 6 403.61 km2 to 7 473.09 km2， and the increase was mainly distributed in the source area of the Yangtze River； the change of the permanent water area of the TRH has a significant positive correlation with the annual precipitation. There is no significant direct correlation between changes in the area of water bodies and temperature； Zhuonai Lake has been dam-breaking since 2011， causing its own shrinkage， leading to serious expansion of the Salt Lake downstream and the risk of dam-breaking.
In the context of "One Belt， One Road"， the current research on the development of built-up areas in border cities is insufficient， especially the expansionary interaction of cities adjacent to domestic and foreign ports needs to be explored. The cities of Ruili and Muse， which are typical border crossing cities in the China-Myanmar border region， are used as the study area. This study selects remote sensing images from 2012， 2015 and 2018， and extracts the built-up areas of the two cities using the Urban Construction Land Index to analyze the dynamic and changing characteristics of the built-up areas of the two cities after the establishment of the Ruili Development and Opening-Up Pilot Zone in terms of the number of expansions， spatial distribution patterns and internal economic scale. Furthermore， the interaction between the evolution of the two cities' built-up areas and the drivers of expansion are also discussed. The results show that under the combined effect of the natural environment， economic industry， policy planning， transportation， ports and other factors， the built-up areas of the two cities have expanded significantly. The expansion of Ruili is most obvious in the western part of the ubran， followed by the east， showing a "one body， two wings" expansion pattern， gradually shifting from the outer extension of built-up areas to the internal filling and infrastructure improvement stage. Muse is expanding mainly to the northwest and due east， and the built-up areas of the city maintain a more pronounced outward character， but the level of intensive land use within the city is low. In the process of interactive and coordinated development， Ruili is dominant and Muse is in a dependent position. The rapid synergistic expansion of the two cities also reflects the close political and economic exchanges and the vast scope for cooperation between China and Myanmar.
Quantifying the United Nations Sustainable Development Goal 11.3.1-"Ratio of Land Consumption Rate to Population Growth Rate （LCRPGR）" is helpful to understand the relationship between urban expansion and population growth， provide data support for urban land space planning and population control， and is crucial to guide decision makers to formulate urban growth plans. Based on the land use products， night lighting data and census data， we extracted the urban built-up areas and used the geographic weighted regression model to mapping the population density of the 1 km × 1 km grid scale in the Pearl River Delta region. Based on the definition and formula in SDG 11.3.1 indicator metadata， the reliable evaluation of the SDG 11.3.1 indicator was achieved in the Pearl River Delta region. The results showed： （1） the built-up area in the Pearl River Delta increased by 4.6 times and the urban population increased by 3.7 times from 1990 to 2010； （2） During the period of 1990~2000 and 2000~2010， the LCRPGR value increased from 0.71 to 2.01. The rate of urban expansion and the rate of population growth were not proportionally coordinated. In summary， the land consumption rate of the Pearl River Delta region has exceeded the population growth rate since 2000. The urban expansion rate and the population growth rate are not proportional. Attention needs to be paid to the rapid expansion of cities.
As an important bamboo species， moso bamboo forests are widely distributed in southern China and has great ecological and economic benefits. However， it is difficult to distinguish moso bamboo forests from other forests. Most of existing extraction methods directly use available clear sky observation， which do not fully consider the influence of classification time phase， limiting the extraction accuracy. Taking Qingyuan county， Zhejiang Province as an example， a method of moso bamboo forest extraction was established in this paper. The characteristics and differences of seasonal spectral curves were evaluated for typical local vegetation types using MODIS high resolution images， and 16 classification experiments were carried out on single and multi-temporal Landsat OLI images. Based on these analysis and experiments， the best seasonal phase to distinguish moso bamboo forest from other vegetation types was selected， and the distribution of moso bamboo forest was extracted effectively by using random forest classifier. The results showed that： （1） Early or middle autumn is the best period to distinguish moso bamboo forest from other vegetation in the study area， followed by summer and worst in winter and spring. （2） When there is no clear-sky observation in early and middle autumn， the extraction accuracy of moso bamboo forest is the best for combination of summer and winter images， with user and producer accuracy of 85.57% and 78.06%， respectively. （3） The extraction accuracy is the highest based on Landsat image in October， with user accuracy and producer accuracy up to 89.00% and 86.91%， and the extraction accuracy is better than 89.23% when compared with the local forestry resources census data. Experiments show that in extraction of moso bamboo forest in similar subtropical areas， the early or middle autumn image should be selected first； if there is no clear-sky observation in this period， the combination of summer and winter images should be chosen priority.
Due to factors such as clouds and atmosphere， there are certain errors in the Normalized Difference Vegetation Index （NDVI） time series data set. The Savitzky-Golay （S-G） filtering method can reduce this error to a certain extent and suppress the sudden drop of low-quality pixel values， but it is lacking in the suppression of high-value low-quality pixels and the protection of high-quality pixel values. ， And cannot be used well in time series image reconstruction at different time intervals. Based on the Google Earth Engine （GEE） cloud platform， a comprehensive use of spatial interpolation， temporal filtering， and pixel quality analysis to reconstruct the 250m resolution MOD13Q1 long-term data set in Chongqing from spring 2014 to winter 2018. At the same time， the Pearson correlation coefficient （Pearson）， the newly proposed smoothness index and the difference of NDVI change are used to quantitatively compare the reconstruction results of a single sample point and a single image. Research shows that under the same parameters， the correlation between the time series reconstructed by the new method and the original image is higher than that of the S-G method； in the simulated noise experiment， the correlation between it and the two simulated noise images are 0.87 and 0.94， respectively， while the correlation of the S-G method is only 0.65 and 0.61.
Forest height is an important parameter in the study of forest ecological and biomass. At present， there are many remote sensing technologies that can obtain tree height， but all have some problems. Populus euphratica in the lower reaches of the Tarim River is the core component and important ecological restoration object of the desert ecosystem in the inland river basin of the arid area.Taking Populus euphratica in the lower reaches of Tarim River as an example， using high-resolution remote sensing image and object-oriented image analysis technology， the canopy of Populus euphratica at single wood scale is obtained， and the corresponding spectral， texture and geometric features are extracted； with support of the tree height data acquired by UAV， the tree height regression model is established by linear， MLP， PACE and SVR respectively. The results show that： （1） the regression model R2 based on the spectral， texture and geometric features is 0.668 7 at the highest， and RMSE is 0.942 6 m， which indicates that the combination of VHR satellite remote sensing and UAV can be used to obtain the height of Populus euphratica at single wood scale； （2） When all features are used， the correlation coefficients of MLP， PACE and SVR regression models are greater than 0.81， and PACE regression archieve the highest accuracy； （3） On the scale of single wood， the spectral features contain more tree height information， followed by texture features.
Urban heat island is a phenomenon that the temperature of urban area is higher than suburb， which can change the natural and social process of city and causes a series of environmental problems. In this paper， Single-channel algorithm for Landsat 8 TIRS10 band （TIRS10_SC algorithm） is used to retrieval the land surface temperature of four landscape Landsat 8 images in Changsha metropolitan area in July 2013， March 2016， July 2016 and November 2016. This paper further analyzes the influence of typical land surfaces such as construction land， green land， rivers and roofs of different materials on the land surface temperature， the results indicate that： （1） The areas with high LST were located in Changsha Railway Station， the Gaoqiao Market and some factories at all times. Compared with July 2013， the heat island effect in the surrounding area of Liuyang River was alleviated in July 2016， which was mainly caused by the different weather conditions and the change of land cover nature by demolition. The largest ratio of construction land in March was moderate LST zone. The highest ratio of construction land in July was sub-high LST zone. In March and July， the highest ratio among green areas is sub-low LST zone， the largest ratio in water is low LST zone. In November， the highest ratio of construction land and green space was medium LST zone， and the sub-highest LST zone in water was the highest； （2） Within 120 m around the river， for every 30 m decrease from land to river， the average temperature of construction land decreased by 0.93~1.26 ℃ and the average temperature of green land decreased by 0.57~0.99 ℃ in July. The average temperature of construction land decreased by 0.51~0.78 ℃ and the average temperature of green land decreased by 0.3~0.57 ℃ in March. The cooling intensity of the river is related to the difference between the river temperature and LST more than 120 m away from the river； （3） Negative MNDWI is positively correlated with land surface temperature and positive MNDWI is negatively correlated with land surface temperature in March and July. However， MNDWI is positively correlated with land surface temperature in November； （4） Emissivity has a significant effect on the results of land surface temperature inversion. It is difficult to distinguish the high reflectivity roofs and other types of construction land by using NDVI to estimate emissivity. Therefore， the influence of high-reflectivity roofs on emissivity needs to be further studied to improve the inversion accuracy of land surface temperature and provide a reference for mitigating the urban heat island effect.
