Building facade is the main content of street images captured by mobile measurement system and contains a lot of regular repeating textures. Applying feature matching algorithm to such images may cause a lot of false matches, which seriously affect the later image orientation and three-dimensional reconstruction. To solve this problem, this paper proposes a phase correlation supported KLT (Kanade-Lucas-Tomasi)feature track-matching algorithm. Firstly, phase correlation algorithm was applied from global to local scale to get crude registration. Then the KLT algorithm was used to track the corners at each matched area. The experimental results show that, when match between building dense street images, the algorithm proposed have a greater increase than pure feature matching algorithm in correct matching rate, and the distribution of features is relatively uniform, which can effectively solve the feature matching problem of street images with regular repeating textures.

In this paper, the authors analyzed the variation law of visible light, thermal infrared band and mid-infrared band of FY geostationary satellite for sea fog, clouds and sea surface (clear sky) in Taiwan Strait, which was based on a lot of experimental analyses by using different phases of satellite data, combined with the visibility data of automatic meteorological stations. On such a basis, reflectivity threshold was used to separate sea fog and cloud from sea surface, and brightness temperature threshold was used to separate sea fog and low cloud from middle and high cloud. In addition, night sea fog was automatically identified by the normalized difference index of mid-infrared and thermal infrared band. Finally, the automatic monitoring software system of Taiwan Strait sea fog was established, and surface observation data were used to examine the precision of remote sensing monitoring. The research results show that FY geostationary satellite could make up for the deficiency of Polar Orbit Satellite in time resolution, and it has a good performance on the dynamic monitoring service of Taiwan Strait. A comparison shows that the remote sensing monitoring results of sea fog are in accordance with observation results, and the monitoring accuracy is more than 70% in daytime. Night time accuracy is lower than that of the day time, and there exists limitation in the separation of sea fog and low cloud.

Leaf area index (LAI) is a primary parameter for characterizing vegetation canopy structure. Since LAI can affect many vegetation ecological processes, such as transpiration, interception and energy exchange, it is used as a critical input for ecological models and land surface process models. At present, several global LAI datasets have been generated from different satellite remote sensing data, such as AVHRR, MODIS and VEGETATION, by different retrieval methods. MODIS, CYCLOPES and GLASS LAI datasets are those with higher spatial and temporal resolution. The spatial and temporal consistency of MODIS, CYCLOPES and GLASS LAI datasets was analyzed over Hanjiang River basin, which is covered with several vegetation types. Comparative study revealed the following characteristics: ① CYCLOPES LAI was observed to contain a large number of missing pixels, while MODIS and GLASS LAI products were more spatially and temporally complete. MODIS LAI contained many invalid pixels, whose LAI became much smaller abruptly in comparison with the LAI values just before or after this time. ② The spatial distributions of MODIS, CYCLOPES and GLASS LAI were mainly consistent with the vegetation types of the basin. The spatial distributions of MODIS and GLASS LAI were more consistent than those of CYCLOPES LAI. MODIS LAI was larger than GLASS LAI in forest pixels, while it was contrary in other pixels. CYCLOPES LAI was much smaller than MODIS and GLASS LAI in forest pixels. ③ MODIS, CYCLOPES and GLASS LAI products generally depicted similar temporal trajectories. GLASS LAI had the smoothest and completest trajectories, while the trajectories of MODIS LAI contained a large number of erratic fluctuations. All of these three LAI products depicted similar seasonal changes for different vegetation types. Compared with CYCLOPES LAI, a good agreement was achieved between MODIS and GLASS LAI values.

The object-oriented method solves the problem of segmentation of objects, divides different features into different objects and to a great extent separates the cultivated land, forest land, water, roads, buildings and other typical objects which are inseparable; nevertheless, the object oriented method for features such as shape, texture description is not comprehensive, the amount of information is not enough to support the whole classification and recognition. In this paper, a new method of combining object-oriented and deep learning is proposed, in which the Caffe framework of convolution neural network is used to study the training sample data in depth and, by grasping the texture of different objects and forming deep learning model, guides the classification of objects. The experiment shows that the new method can effectively solve the problem of the low classification accuracy.

In order to realize the automatic change detection with vector map and remote sensing imagery, a change detection method based on the object heterogeneity for vector map and remote sensing imagery is proposed in the paper. Image segmentation under the constraint of vector map was employed to get image objects using marker-based watershed algorithm. The features of the object were extracted by histogram which describes both gray feature and texture feature. The histogram intersection distance was adopted to measure the feature distance. The object heterogeneity was built by the average of the distance between the object and the other objects with the same class in old period. Change/nochange label of the objects can be determined by comparison the object heterogeneity with the heterogeneity threshold of the class which was calculated by Maximum Entropy Principle automatically. Experiments on QuickBird remote sensing images verified the effectiveness of the proposed method ,and the correct rate of the change detection is up to 95%.

In view of the problem that the traditional super-pixel Markov random field (MRF) image segmentation model cannot fully utilize spatial context information, a new super-pixel MRF model is proposed. This algorithm incorporates higher-order neighborhood model into the interactive potential term of MRF. The new model enables the interactive potential to fully exploit the spatial context information contained in the super-pixel neighborhood system. Additionally, a new class-wise estimation method for β is proposed, which is based on norm distance. By utilizing two scenes of high-resolution remote sensing images acquired over different agricultural landscapes, validation experiment was conducted. The experiment results indicate that the proposed method can better use the contextual information such as edge strength, thus achieving higher segmentation accuracy. Moreover, the algorithm proposed by the authors showed superior performance when it was compared with other super-pixel MRF approaches.

Two mono-window algorithms, i.e., radiation transfer model (RTM) and QK&B algorithm, were evaluated and compared for their performance on sea surface temperature (SST) calculation from Landsat8/TIRS data in coastal water of Hongyan River nuclear power station. The parameters of QK&B algorithm were modified for Landsat8 thermal infrared band 10 based on thermodynamic initial guess retrieval (TIGR) atmospheric profile data, and both atmospheric transmittance values obtained from water vapor content based on empirical model and from National Centers for Environmental Prediction (NCEP) data were employed for QK&B algorithm respectively with the purpose of comparing their feasibilities for SST retrieval. A validation with shipboard measurements of SST collected synchronically shows that the slightly better accuracy in SST retrieval is observed from RTM method (RMSE = 0.525 0) than from modified QK&B algorithm (RMSE = 0.638 0). QK&B algorithm using atmospheric transmittance simulated using NCEP data provided better accuracy than that using atmospheric transmittance estimated from water vapor content. Sensitivity analysis based on data simulated by MODTRAN4.0 using NCEP data was conducted. The results show that atmospheric transmittance has the greatest impact on the accuracy of SST retrieval among all parameters input RTM method, followed by atmospheric upward radiation. Atmospheric transmittance also shows greater sensitivity for RTM method than that for QK&B algorithm. For QK&B algorithm, atmospheric average temperature has greater impact on SST retrieval than other parameters input. Atmospheric upward radiation for RTM method, atmospheric average temperature for QK&B algorithm and atmospheric transmittance for both two algorithms show increasing impact on SST retrieval with increasing water vapor content.

The application of the high spatio-temporal resolution data possesses very extensive foreground. Consequently, based on a flexible spatio-temporal data fusion(FSDAF)method and using MODIS and ASTER data,the authors generate the land surface temperature(LST) with high spatial and temporal resolution. FSDAF is a method based on spectral unmixing and thin plate spline interpolation function. Compared with the existing spatio-temporal data fusion method, its advantages lie in less input data,suitableness for heterogeneous surface and capability of predicting the gradient of land cover types and so on. The fusion results were verified by using the ASTER temperature products(7 days) and the surface radiation infrared temperature data(4 days)of the automatic weather station(AWS) sites. The results show that the LST images generated by the data fusion method based on FSDAF have higher clarity, the correlation coefficient of the fusion images and the ASTER LST products is higher than 0.91(September 28) , the room mean square error (RMSE) is less than 2.44 k(September 19), the mean absolute error (MAE) is less than 1.84 k (September 19)and the correlation coefficient of the fusion images and the AWS LST data R2 is higher than 0.64(August 18).

Detecting snow cover information and snow space-time distribution quickly and accurately is a basic problem of ecological environment changes in the resources. Remote sensing technology effectively provides technical support for solving this problem. Normalized difference snow index (NDSI) is an important method for automatic extracting snow cover information using spectral features of snow, which have high reflection in the green band (0.53~0.59 μm) and strong absorption characteristics in short wave infrared band (1.57~1.65 μm). By using Landsat8 OLI images as the data source and according to the spectral characteristics of snow, the authors propose the enhanced normalized difference snow index (ENDSI) based on adding emissivity characteristics of snow in first band B1 (0.433~0.453 μm) and second band B2 (0.450~0.515 μm), and the utilization of this index to extract snow from OLI images. Simulation and case study results show the following characteristics: the sensitivity of ENDSI is stronger than that of NDSI for the snow thickness; with the increase of the thickness of snow, the change of ENDSI value is stronger than that of NDSI; ENDSI can effectively increase the difference between snow and non-snow; it is easy to extract snow from the image with 0.3 as ENDSI threshold and, in this way, snow extraction accuracy is improved.

Supervised classification methods usually require adequate labeled samples which are difficult and time-consuming to obtain for polarimetric SAR images, while the expression capability of the shallow structure learning algorithm is limited. A novel supervised classification method for polarimetric SAR imagery based on active deep learning is proposed in this paper. Firstly, the features are extracted from an original image by multiple polarization target decomposition methods for fully describing the data,and the features which are separable and invariable can be extracted with unsupervised learning by auto-encoder. Then, the initial classifier is trained and fine-tune the whole model with a small number of labeled samples. Finally, the most valuable samples (the largest ambiguity samples for classifier)are selected to label by active learning. Experimental results in comparison with conventional methods for polarimetric SAR data sets of RADARSAT-2 and EMISAR show that the proposed method can achieve higher classification accuracy with a small number of labeled samples.

Land surface temperature(LST)is an important parameter in the model of energy balance of the earth surface. The enhanced spatial resolution of high temporal resolution of remote sensing surface temperature can be realized by downscaling algorithm, which is of great significance for monitoring the spatial and temporal distribution of the LST. In this paper, Beijing City was taken as the study area, and the LST with 100 m spatial resolution was retrieved by using Landsat8 OLI/TIRS data through improved mono-window(IMW)algorithm,which was used as validation data. Besides,the normalized difference vegetation index(NDVI),normalized difference built-up index(NDBI)and other remote sensing index were calculated and simulated to the spatial resolution of 1 000 m, which was united with the MODIS/LST with the spatial resolution of 1 000 m to be input into the random forest(RF)model to acquire downscaled LST(100 m). Meanwhile, the downscaled results retrieved by RF algorithm were compared with the two commonly used methods of downscaling, multi factor regression method and LST sharpening algorithm based on vegetation index (TsHARP). The results show that, with the simulated Landsat/LST as the data source, the RMSE of downscaling LST retrieved by RF was 2.01 K, and the RMSE was improved by 0.16 K and 0.44 K compared with the multi factor regression method and TsHARP algorithm respectively. For the MODIS/LST, the RMSE of downscaling LST retrieved by RF was 2.29 K, and the RMSE was improved by 0.42 K and 0.50 K compared with multi factor regression method and TsHARP algorithm respectively. For different land surface types, the effects of RF downscaling algorithm are different. The effect of high vegetation coverage area is the best, and the RMSE is 1.81 K. Due to the spatial heterogeneity of the urban surface, the RMSEhas reached a maximum of 2.75 K.

The purpose of this paper is to improve the precision of salinity monitoring model with Landsat8 OLI multi-spectral images in the oasis of arid area. In this paper, the authors chose the Ebinur Lake region as the study area, and reflectivity of saline soil based on OLI image and spectral reflectivity from resampled ASD data were measured respectively. According to the findings of the correlation analysis of twelve transforms of soil spectral reflectance with soil salt content, multiple stepwise regression analysis algorithm was used. Based on the analysis, the authors chose the most sensitive band ranges to establish a soil salinization monitoring model using the ASD actual measurement data and corrected OLI image inversion of soil salinity. The results show that the soil salt content inversion model based on the measured field spectral is satisfying, the first-order of the logarithm of the reciprocal with the best accuracy and the R² is 0.779. Spectral reflectivity after resampling data performed better than those monitoring models with OLI spectral data, the coefficient of determination (R²) is raised from 0.28 to 0.777 6, and the RMSE is 0.281. The authors realized the scale transformation of the soil salt content spectral inversion model from field measurements of spectral scales to spectral scale of multi-spectral remote sensing, and the results could provide a theoretical reference for further improvement of the accuracy of quantitative remote sensing monitoring of soil salt content at the regional scale.

In order to study the methods for forest cover change monitoring by using GF-1 images, the Yajiang County in Sichuan Province was selected as the research area to extract the information of forest coverage based on the two GF-1 WFV data. Firstly, the data were normalized by using the iteration re-weight multivariate alteration detection(IR-MAD)method. The two images were transformed by kernel principal component analysis(KPCA)method, and formed differencing image. Then, the changed area was extracted using the method of maximum between class variances(OTSU)for automatic threshold selection. Finally, the change detection results were validated using OTSU with the field sample data, and the extracted results were verified by way of precision test, and comparatively analyzed with the change vector analysis(CVA). The research results show that the overall accuracy of the two change detection methods is higher than 80%, and the overall accuracy of the KPCA method is 89.27%. The user precision of unchanged area is 93.88%, and the user's accuracy of changed area is 80.28%. The accuracy of the KPCA method is better than that of the algorithm based on the traditional CVA method. It is shown that, after the data transformation, KPCA algorithm can reduce the correlation between the variables and enhance the signal to noise ratio of the image, thus improving the recognition accuracy for the changed area.

In this paper, Yongding District, where there are World Heritage Hakka Tulou and its Conservation Plan Area(Hugao region), were taken as the study area. With the purpose of analyzing the land cover and ecological environment changes in the study area from 1988 to 2014, the authors selected five Landsat images acquired in 1988, 1996, 2002, 2009, 2014 respectively to generate the land cover classification maps by using multiple classifier system and the remote sensing based ecological index (RSEI)maps by using RSEI. The change of land cover and ecological environment in the study area was obtained by the method of post classification change detection. Some conclusions have been reached: ① The land cover of the study area significantly changed during the 26 years from 1988 to 2014, and the forest, shrub and grassland and built-up land increased rapidly, while farmland, degraded land and cultivation land decreased considerably; ② RSEI is suitable for the evaluation of the ecological environment of the World Heritage Hakka Tulou; ③ The overall ecological quality of the two study areas improved year by year except for the period of 1988―1996; ④ The relationship between the land cover and the ecological environment revealed that the ecological quality improved in the degraded land at high-altitude,the slope farmland which was converted into the forest land, shrub and grassland, and became worse in the area of forest, shrub and grassland land which degraded into the area of farmland, the degraded land and the reclaimed land as well as the expansion area surrounding the urban district.

In this paper, the authors aim to clarify the spatio-temporal patterns of cloud interference in Dongjiang River basin and then provide scientific basis for vegetation index studies at the level of higher precision in this area. Cloud flag information of MOD13Q1 product data sets within the range of Dongjiang River basin from 2001 to 2015 were analyzed using the GIS approach. Three aspects were analyzed, i.e., the probability of cloud interference and its spatial distribution, the elimination rate of cloud interference and its spatial distribution and the seasonal differences of cloud interference duration period length. The results show that the whole probability of cloud interference was reduced rapidly with the increase of compounding period length, the new added elimination rate of cloud interference firstly increased and then decreased with the increase of compounding period length. The duration period length of cloud interference was relatively longer in southern urbanized region, and in summer and spring. The results of this study can provide scientific basis for choosing or developing better methods of removing clouds for vegetation index time-series data.

The remote sensing images which were obtained respectively in 2008 and 2014 were used in a certain area of Shanxi Province. By using ArcGIS softwere, human and computer interaction interpretation method was used to delineate the mining land and non-mining land respectively. The monitoring results show that, from 2008 to 2014, the proportion of mining land in the study area increased by about 35%, and the mining lands grew rapidly. The change of mining manner was the main reason for the increase of mining land. The increase of mining land was mainly attributed to the occupation of the forest land and cultivated land.

Remote sensing image can reflect distribution rules and structural features of active faults exhibition space from the macro. Based on regional geological study, the authors used data preprocessing, information enhancement and data fusion to improve the degree of clarity and interpretation of remote sensing images. The interpretation keys of the fault zone were established based on the spectral characteristics and geometric flag of the remote sensing images, and the interpretation of the faults and their activities was carried out. Combined with field investigation, the macro analysis and segmentation activity study of Macheng ―Tuanfeng fault were carried out. The results show that Macheng - Tuanfeng fault zone can be divided into northern, middle and southern sections according to the intensity of the control force. The linear characteristics of the northern section are apparent in the image, and the interpretation key of the fault is remarkable. The fracture control force of the middle section is weak, whereas the linear characteristics of the image are fuzzy. The southern section is a buried fault. Remote sensing technology plays a very important role in the survey of the activity of Macheng―Tuanfeng fault zone; the application of high resolution remote sensing imagery and remote sensing image processing technology, in particular, can not only speed up the progress of the investigation but also provide guiding information for the field survey, so as to improve the survey efficiency and accuracy.

With rare metals of granite pegmatite type in the Dahongliutan area of West Kunlun Mountains as the study object and WorldView-2 remote sensing images as the major data source, the authors drew standard image map, adopted methods of image enhancement for protruding the information of ore-controlling factors and mineralization, and finally carried out an interpretation of remote sensing for mineral resources. On the basis of alteration anomaly information extraction for rare metals of granite pegmatite type by using ASTER data and a right amount of field verification, its high-resolution remote sensing characteristics and metallogenic geological conditions were analyzed, and a remote sensing geological prospecting model was established. Tt can provide the basis for finding similar minerals in the West Kunlun metallogenic belt in future. The results show that the remote sensing technology using high resolution satellite images can be used as an effective method for detection of potential mineral resources enrichment region,which can meet the requirements of rare metal mineralization belt resources exploration and assessment in the Dahongliutan area of West Kunlun Mountains and hence deserves further promotion and application in the same area or similar areas.

To study the spatio-temporal variation of urban heat island effects in Fangchenggang City from 2001 to 2015,the authors used the remote sensing methods to monitor the variation characteristics of urban heat island effects in Fangchenggang City for a period of 15 years. The land surface temperature(LST)was retrieved using remote sensing images(Landsat5 TM and Landsat8 OLI) acquired in three periods of 2001, 2008 and 2015. Then, both urban heat island intensity and urban-heat-island ratio index were constructed to analyze the evolution characteristics of heat island effect in the past 15 years from three aspects: the space-time distribution and area variation of heat island intensity, the development characteristics of urban-heat-island ratio index and the influence of underlying surface condition on heat island effect. Some conclusions have been reached: ① Urban area exhibits a trend of rapid expansion in the study area. ② The urban heat island intensity increases year by year, especially in Gangkou District, where annual growth rate reaches 26.72%. ③ The urban-heat-island ratio index is rising year by year in all districts, among which, Dongxing reaches the highest value of 0.62. ④ Cooling effects are obviously for both urban green space and water body, but the operating distance and cooling amplitude of water body are larger than those of green space. However, the proportion of urban vegetation and water of the study area is markedly low. The research results may provide scientific and reasonable proposals for Fangchenggang government's aim of reaching the goal of creating a national garden city.

The purpose of this paper is to study the potential application of ZY1-02C satellite data when they are used in the checkup of land consolidation project. The land consolidation project in Neijiang of Sichuan Province was selected as a case study for this research. Based on the software including ERDAS, AutoCAD and ArcGIS, the authors carried out the indoor verification and the field investigation to put forward the complete technical method and operation procedures by using 02C satellite data for the check up of land consolidation project. The results show that the 02C satellite images could be used to evaluate the cultivation in land leveling project, the shanping pool in irrigation and drainage engineering, and the field road in road project. Nevertheless, the projects, such as the bank of paddy field, field drain, water tank and the passing track, can hardly be checked. The quality of project cannot be evaluated. The conclusion of the paper is that the 02C satellite image can be used in land consolidation project checkup.

In the period of “13th Five-Year Plan”, the national planning attaches great importance to the implementation of construction of the high standard grain area. As one of the high standard grain areas in Henan Province, Hebi City shoulders the burden of ensuring food security. That is why it is of practical significance to study the simulation and prediction of land use in Hebi City in the future. With the use of CA-Markov model and on the basis of the historical process of the coordinated development of urbanization, industrialization and agricultural modernization of Henan Province, the prediction can be divided into two scenarios for the simulation and prediction of land use according to the analysis of the characteristics of land use in Hebi city in the past 20 years. The results can show that the Kappa index of Hebi City in 2013 was about 0.898, which means that the fitting effect is the best, and that the prediction results of CA-Markov model can achieve good fitting effect. Based on comparative analysis of the quantity, the space and the landscape index of Hebi City, it is held that the scenario II can be more in line with the demand of the Central Plains Economic Area and the industrial development as well as with the “green development” in the High Standard Grain Area’s ecological and environmental protection grounds. What’s more, it is in accordance with the planning of Hebi City, the patch shape is more regular, the plaque agglomeration degree is high, the internal continuity is strong, landscape fragmentation degree is low, and the landscape distribution is uniform, thus exhibiting obvious advantages. The authors hold that, in the future, the government should adhere to scenario II model for the development of the construction of the High Standard Grain Area so that the land can realize sustainable development.

Describing, quantifying and monitoring land use/land cover change play an important role in the global and environmental change investigation. Based on a comprehensive change detection method (CCDM), the authors mapped the land cover changes in the period between 2011 and 2016 over the Landsat image (Path33/Row33). CCDM integrates some spectral change based detection algorithms, which encompasses the multi-index integrated change analysis (MIICA) model and the Zone model, with an emphasis or core on MIICA. By calculating four spectral indices including change vector (CV), relative change vector maximum (RCVMAX), differenced normalized burn ratio (dNBR) and differenced normalized difference vegetation index (dNDVI), the land cover changes were extracted from the bi-temporal imagery. According to the previous and the current land cover change trends, coupled with the changes in results from MIICA and Zone models, the accuracies of change detection results by CCDM were evaluated. The results show that there is an accuracy of 96% for the category of no-change, and 40% for change category, with an overall accuracy of 68%. The CCDM is a simple, easily realized and widely used model to capture the potential land cover changes caused by the diverse natural and anthropengic disturbances in different landscapes.

In the past decade, Ordos City of Inner Mongolia experienced rapid development of economy and rapid urban expansion, and now it is facing the stagnation of economic development and urban transformation. Urban expansion mapping and driving factor analysis at temporal and spatial scales could provide valuable information for the urbanization development planning of Ordos. Using the DMSP/OLS nighttime light(NTL)data combined with Landsat image, the authors compiled urban expansion maps for the period from 2000 to 2013. The build-up areas were extracted using a radiometric correction method based on the stable point NTLs, support vector machine (SVM) classification and thresholding techniques. The statistical analysis results show that the expansion of Ordos City can be mainly divided into three stages, and the driving factors of urban expansion could be quantitatively analyzed using the portfolio analysis method.

Rape is the major oil crop in China, and timely and accurately obtaining the information about rape planting spatial pattern has great significance for growth monitoring, yield estimation and disaster assessment. In this paper, the distribution of rape growing in the Jianghan Plain from 2014 to 2015 was extracted utilizing the MODIS EVI time series data with 250 m spatial resolution. The field survey data were used to extract crop training samples for MODIS EVI data indirectly by using TM data as the transition data between the field survey data and MODIS EVI image. According to the spectra and phenological calendar of winter wheat and rape in Jianghan Plain, the authors established the extraction model for the area of rape growing by multiple threshold comparative method. With the Agricultural Bureau Statistics data as the verification, the overall accuracy of the extraction results of MODIS data were up to 95.22% and 91.29% respectively in 2014 and 2015. In addition, the extraction result was quite consistent with TM-based result with a precision of 88.61%, in 2014. The results show that, based on the time series MODIS-EVI data sets, combined with the EVI spectral characteristics and phenological information of crop and using the study method presented in this paper, the rape planting distribution information could be extracted effectively in Jianghan Plain.

Henan Province was therefore chosen as the study area in this paper. The temperature condition index (TCI),vegetation condition index (VCI) and rainfall condition index (TRCI) were calculated by using the EOS-MODIS surface temperature product MOD11A2,vegetation index products MOD13A3 and tropical rainfall measuring mission’s monthly rainfall rate data TRMM3B43 from 2005 to 2014. Then an agricultural drought index (ADI) monitoring model was built up by weighing these three indices and the weights were determined by using analytic hierarchy process (AHP) analysis method. Grading indices of ADI were derived in terms of the classification scales of drought for standardized precipitation index (SPI) which were authorized by national standard. Finally,the ADI monitoring model was constructed to analyze the drought situation in Henan Province in 2014. The results indicate that the ADI proposed in this paper can well capture the temporal and spatial characteristics of the drought situation in Henan Province.

The Dananhu district in Hami of Xinjiang is a typical gobi desert in Northwest China. In this paper, the authors investigated the temporal-spatial variations of the natural vegetation coverage during 1992―2014 in this region, using correlation analyses and dimidiate pixel model based on the multi-spectral remote sensing data, the local meteorological data, and the digital elevation model. The results show that, from 1992 to 2014, vegetation coverage in the region showed a trend of increase. Generally the vegetation coverage is weakly positively correlated to elevation; locally, however, the plants (mainly juniper tamarisk, haloxylon ammodendron, and reed) are more developed in the relatively depressed localities (saline areas or sandy dry riverbeds) than those in Gobi desert areas. The vegetation coverage is positively correlated to the sunshine duration and evaporation, but unrelated to precipitation and humidity. It is suggested that the natural plants in this regions live on groundwater mainly. The global temperature increasing during 1992-2014 might to some extent promote, instead of retard, the natural vegetation, probably through enhancing the groundwater supply due to glacier melting at nearby mountains.

Leaf area index (LAI) is an important agricultural parameter to assess crop growing status for production estimation. Due to its very high spatial resolution, GF-2 can be used as a new source of remote sensing data for crop monitoring. It is particularly valuable to develop approaches for LAI estimation using GF-2 data. In this paper, the study of LAI estimation for wheat at the booting stage in Wanzhuang Town, Langfang City of Hebei Province in North China Plain is presented. Canopy LAI of wheat at the booting stage was measured over an experimental field in the town. Regression analysis method was performed with LAI and four different vegetation indexes, and the neural network method combined with the PROSAIL model was also considered. The results show that the best LAI retrieval regression model is the binomial model of normalized difference vegetation index (NDVI). The correlation coefficient (R²) and root mean square error (RMSE) are 0.719 3 and 0.393 6 respectively. Compared with the regression analysis method, the accuracy is greatly improved by using neural network method,and the R² and RMSE reached 0.900 8 and 0.273 2 respectively. Neural network method has feasibility and applicability based GF-2 data at wheat booting stage, which would provide a reference scheme for LAI inversion from high spatial resolution satellite image.

Using remote sensing method to study eco-environment quality changes is immediate and rapid. Based on remote sensing image and principle component analysis, the authors combined vegetation index, wet index, dryness index and temperature index to evaluate the eco-environment quality of Baoji City and the land-use types from 2002 to 2013. The results indicate that the ecological construction work has made remarkable achievements in the past 10 years. The area of improved eco-environment quality reached 39.50%, while the area of degraded quality possessed only 10.96%; in addition, synthetical ecological index (ESI) increased from 3.25 to 3.56. The eco-environment quality of Chencang District and Weibin District was improved while that of Jintai District was degraded. The ESI of land-use types from high to low is forest land, unexploited land, grassland, waters, cultivated land and construction land. RSEI degree of cultivated land has no significant change, that of forest land, grassland and unexploited land have been improved while water and construction land have been degraded.

Monitoring the planting area and growth status of rare medical herbs is a new direction of remote sensing application. Pseudo-ginseng, as a commercial and severe planted crop in Yunnan Province, has a significant influence on the regional land use and financial income. Thus, the accurate monitoring of the planting area and the trend of pseudo-ginseng is of vital importance. Based on the analysis of spectral and spatial features of the pseudo-ginseng plantation, an efficient method to extract the planting areas of pseudo-ginseng was proposed and its change was analyzed using multi-temporal remote sensing images. Experimental results demonstrated that the planting area of pseudo-ginseng in Wenshan and Honghe has kept grown from 2010 to 2015, especially from 2012 to 2014 with a rapid speed. However, the market price decreased because supply exceeds demand in this period. Compared to the market price, the planting area has a obvious lag effect. Furthermore, the pseudo-ginseng plantation is concentrated in the appropriate elevation gradients between 1400 and 2000 m, with the slope below 25°.

With the Manas Lake in Xinjiang as the study area,by using eight remote sensing images from 1972 to 2014, and on the basis of extracting the information of lake water, the authors analyzed the evolution of the Manas Lake area in the past forty years. The results are as follows: The Manas Lake showed a significant “increase-decrease-increase-decrease” trend of change from 1972 to 2014, and the barycenter of water area was migrating in southwest direction. Lake had dried up from 1972 to 1999, then it restored the largest water area (248.69 km²) in 2000,and after that it experienced two shrinking periods that happened in 2000—2008 and 2011—2014 respectively. Calculation result of area variation amplitude and dynamic degree shows that the shrinking period of the water area was shortened and the shrinking speed increased. In the past forty years or so, the change trend of lake water area was not consistent with the change trend of the Manas River. The authors point out that climate change has little effect on the change of the water area of the Manas Lake, but has an intimate relationship with extraordinary flood caused by extremely high temperature and extreme precipitation, and that human activities in the basin constitute the main reason for the evolution of the Manas Lake, which causes the declining of the function of supplying water resources to the lower reaches.

A severe haze pollution occurred in the Yangtze River Delta during December 1 st-9 th, 2013. The formation, characteristics and potential sources of this haze pollution were explored by using the aerosol products of CALIPSO/CALIOP and COMS/GOCI satellite data, CE318 sun photometer, ground meteorological data and HYSPLIT backward trajectory model. The six typical cities of the Yangtze River Delta were dominated by this haze pollution. The air quality was in severe pollution and the air quality index (AQI) and the value of PM2.5 were higher than those of usual case. The analysis of CALIPSO/CALIOP satellite data indicated that the aerosol mainly existed in the level from the ground to 2 km above, the proportion of spherical aerosol in haze aerosol was higher than that of non-spherical aerosol, and the proportion of large size aerosol particles in haze aerosol was higher than that of small aerosol particles. The analysis results of the CE318 data and HYSPLIT model show that this haze pollution was not only influenced by the local aerosol emissions but also influenced by the outer sources of aerosols, and such factors as continuous low wind speed,high relative humidity and no precipitation during the haze pollution led to the widespread, long-time and severe pollution event in the Yangtze River Delta.

To study the spatial distribution of population density in illegal buildings in Shenzhen, the authors used GIS technologies such as spatial autocorrelation analysis method and regression analysis method. Spatial autocorrelation analysis method was used to study the spatial distribution pattern of the population density in illegal buildings in Shenzhen. Both monocentric population density models and multi-centric population density models were fitted by regression analysis method. The study results show that the spatial clustering pattern of the population density is existent according to the global Moran' s I index, and the area joining the Futian, Luohu and Longhua district is the hot spot area according to the Local Moran’s I and Getis-Ord Gi* index. The fitting results show that, among all the population density models, the quadratic exponential model can best describe population density distribution in illegal buildings of Shenzhen, while the overall fitting results of multi-centric models are at low level, which shows that the multiple centers fail to form the population in illegal buildings. The quadratic exponential model can not only obtain the best result from statistical point of view but also verify that it fits the crator theory well with rent data of illegal buildings.

Fast spatial query is the foundation of the mass ZY1-02C data application. In the traditional spatial query structure using relational database, frequent disk I/Os and data swaps degrade the query performance a lot. In contrast, main memory database which is completely based on the memory can significantly improve the searching efficiency. In this study, a vector data storage and R-tree indexing structure was designed and implemented based on the Redis database, which is a main memory database of Key-Value type. This structure was used in practice and the efficiency of spatial query was improved.