Nowadays,vast amount of remote sensing data have been acquired with the rapid development of Earth Observation System(EOS).It has become a serious task to manage and use these data for most RS and GIS applications.The content-based retrieval system for remote sensing images(CBRSIR) has become resultingly a hot research field with the potential to retrieval interesting information from image databases automatically and intelligently.In this work,we put forward a new remote sensing image retrieval approach by using multi-features including image color and texture.Firstly,a given image is processed by principal components analysis and then decomposed by Quin-tree,which splits large-scale remote sensing imagery into sub images.Secondly,texture features of each image block are extracted via multi-channel Gabor filters,and the standard deviation and third moment of each sub image are extracted as color features.Then,color and texture histograms are constructed based on sub images.Finally,we compare the similarity of the color and texture histograms between the query example and each one in the image database.If the total similarity is higher than some threshold,the image will be returned.These images are sorted according their similarity as the final retrieval results.This approach is validated using high resolution remote sensing images.

The influenza A(H1N1) pandemic was root in the gene mutation of swine,so it is also called "swine influenza".Since the introduction of the first case of H1N1 in China,the epidemic had been rapidly spreading over most provinces of the country in several months and would present a situation of large broken out.The paper analyzed the pandemic H1N1 influenza A in Shenzhen,China during May 26,2009 to November 15,2009.The data of cases rooted in the direct network reporting system for infectious diseases.First,the statistic cases were analyzed according to gender,age and vocation of the sufferers,which shows that the age of sufferers is mainly clustered in the rank of 10-20,more than half of the total cases,and the vocation which has the most infectors is student.Then,temporal distribution in the unit of daily cases was analyzed.The morbidity had been in a relative low level in the beginning 3 months.It suddenly increased rapidly in September and reached the peak on September 6th 2009.That's because early September is the time that students got back to school,the mutual infect among students led to the rapid increase of H1N1 influenza A cases.Finally,the data of cases was processed by GIS.The sufferers' home addresses were selected as the basic units of geo-coding and google earth tool was used to find the longitude and latitude of each sufferer.The retrospective space-time permutation scan statistic was employed to detect the space-time clusters.The longest incubation period of 7 days was chosen as the maximum temporal cluster size and the maximum spatial cluster size is defined 10 kilometers,approximately equal to the north-south distance of central Shenzhen city.The result shows that the space-time clusters of H1N1 influenza A in Shenzhen City are mainly in the northern areas which border with Hong Kong during early September.The come-and-go between Shenzhen and Hong Kong are very frequent,which caused that the epidemic in Hong Kong affect that in Shenzhen a lot.Thus,the government should enhance the control of virus in the customs between Shenzhen and Hong Kong.

Leaf area index(LAI) is an essential parameter for monitoring crop growth dynamic.In this paper,an algorithm,which is based on physical model and neural networks to derive leaf area index from land surface reflectance data,is presented.The algorithm utilizes MODIS land surface reflectances and 4-scale model to produce crop LAI.Firstly,the training dataset was created by running the 4-scale model to simulate crop LAI at different land surface reflectance and geometric situation.Then,the neural network was trained with the simulated LAI data set.After the neural network was trained,the LAI would be efficiently retrieved from MODIS reflectances and geometric data.This algorithm directly utilizes the directional reflectances instead of the BRDF normalized data in order to avoid complex BRDF normalization and the error from it.The estimated LAI is compared with existing LAI products.The results show that it is consistent with MODIS(RMSE=0.4994) and CYCLOPES(RMSE=0.6658) LAI products in temporal and spatial patterns.The algorithm is validated against ground measurements of annual crop LAI of 2004 in Hengshui,Hebei Province,China.The neural network derived LAI could represent the spatial pattern of the field LAI.However,all these LAI products are lower than field measurements.It would be suggested that the physical model should be modified to adapt to the dense crop in Northern China.

The fast expansion of urban built-up land and accompanied sharp decrease in farm land have made timely monitoring of landuse changes become more important than ever before.The ability to monitor the built-up land dynamics in a cost-effective manner is highly desirable for local communities and decision makers alike.Fortunately,satellite remote sensing technique offers considerable promise to meet this requirement.Although the use of remote sensing technique in the monitoring of land use changes has become more and more popular and satellite imagery has been frequently used to discriminate built-up lands from non-built-up lands for the last few decades,the extraction of built-up land information from remote sensing imagery is still not an easy task due largely to the heterogeneous characteristics of the built-up land.Among many techniques developed for the extraction of built-up land information,the index-based built-up index(IBI) was created based on three existing thematic indices rather than original multispectral bands.The use of the three thematic indices-soil-adjusted vegetation index(SAVI),modified normalized difference water index(MNDWI) and normalized difference built-up index(NDBI)-greatly help the delineation of built-up land features in remote sensing imagery,because these three indices represent three major landuse components,which are vegetation,water and built-up land,respectively.Therefore,the IBI can significantly enhance built-up land information while suppressing background noise.Consequently,the built-up land can be effectively extracted from the IBI image with high accuracy.In order to quicken image processing,this built-up extraction technique has been programmed to form an easy-use module using the ER Mapper scripting language.The module was further integrated in the ER Mapper package by adding a button to the manual bar.This allows users to automatically perform the extraction procedure with high accuracy just in a few minutes.

Geographic data collection is very basal and important for GIS.Some specialized data,such as cases,warning instances and so on,can't be collected by remote sensing,map digitization or other general geographic data acquisition method,and many public security service information collection methods about spatial locations don't require the accuracy as the same as the professional surveying and mapping.So,these data are collected usually through labeling on electronic map manually.However,the general manual collecting methods manage the geographic information independently,lacking of correlation and inheritance.Because public security service information and the management method about these data are multiform,and there are high correlation and strong artificially specified relationship between many data,research on a new manual acquisition mode about geographic data is particularly necessary.According to the business characteristics and the acquisition mode,public security service information is divided into independent information and information with associated data.So the concept about main layer and sub-layer is presented,and a method about geographic information association acquisition based on the electronic map manually is researched.Not only can this method ensure many geographic data which have the same spatial location are the same coordinate in the memory and the inheritance of the properties,but also get certain accuracy and realize specified association between many geographic data.This association acquisition method can greatly reduce the workload in the manual data acquisition,make the association relationship between the data more clear and accurate,and meet the geographic data collection,management needs in public security bureau.

In the academic domain of remote sensing information extraction,per-pixel spectral classification and object-oriented classification method are two main approaches.Presently,the application scope of object-oriented classification method is becoming enlarged so as to meet various needs of image processing and information extraction towards diverse ground objectives.The object-oriented method has been a useful helper in many cases;unfortunately it has not developed into a great approach.In fact,several problems have been spotted when applying object-oriented methods.This paper intends to deal with the problems in executing object-oriented information extraction and proposes a new concept,which is to combine the per-pixel classification and object-oriented classification,based on above mentioned delicate problem-studying.The methodology of this paper is presented as follows.To begin with,the priori regional knowledge is first obtained by means of geographic analysis based on traditional per-pixel maximum likelihood classification.By doing this,the spectral confusion between different types of ground objects and the complexity of the whole image can be described both quantitatively and qualitatively,in order to reflect the difficulty and technological requirements regarding image processing and information extraction.Subsequently,the knowledge gained via pre-analysis is then used to direct the technique taches in image processing,as well as to optimize the segmentation parameters and the classifier construction in the workflow of object-oriented extraction.A major objective of the abovementioned technique is to conquer the blindness of parameters setting and the multiformity of segmentation results.Finally,compared to the object-oriented result without the support from geographic understanding and analysis,this proposed method is proved good at partitioning the real elementary objects and improving the classification accuracy,efficiency and the training performance.

The space-time changes of landscape pattern in Changsha-Zhuzhou-Xiangtan Metropolitan Region have been analyzed in terms of Landsat TM/ETM+ images from 1989,1996,2000 and 2008 based on GIS and RS techniques,with its driving forces discussed.The principle and technique of landscape ecology were used to study the landscape pattern change.Results show that there was a remarkable difference in the area change of various landscape types in nearly 20 years between 1989 and 2008 in Changsha-Zhuzhou-Xiangtan Metropolitan Region.The main change in landscape area is that: the farmland and woodland were changed into built-up area.The landscape types' change rate increased gradually from 1989 to 2008.Landscape change is also obviously in the decades.Totally,landscape number and landscape density increases,the largest patch index decreases gradual,the patch becomes more complexity.From the change of Shannon Diversity Index and Shannon Evenness Index,we know that the diversity of landscape and the Interspersion Juxtaposition Index increase.The patch shape becomes more complex with the increasing landscape shape index.The driving force analysis of landscape pattern dynamics shows that three main human factors played an important role in driving landscape pattern changes in Changsha-Zhuzhou-Xiangtan Metropolitan Region.They are population growth,economic development and guiding effect of policies.

Wetland vegetation is one of the important components of wetland ecosystems,and its biomass is the key indicator of the health status of the wetland ecosystems.The sensitivity of near-infrared band to biomass decreases as the vegetation density increases.In contrast,there is no significant change in red spectral reflectivity,so the vegetation index cannot reflect changes in high-density vegetation area.When using C-band SAR data to estimate the biomass in areas with low-and middle-biomass,the radar backscattering from the wetland soil with a great deal of water behind the canopy results in the phenomenon that different biomass have the similar total radar scattering coefficient in the radar images.In this paper,we have done research on the sensitivity to the biomass for three kinds of optical remote sensing vegetation indices NDVI,RVI and DVI,the use of improved MIMICS model to simulate different scattering components of wetland vegetation,and on seting up the simulation database of all components of wetland vegetation backscattering.Then we offer a method with Landsat TM and ENVISAT ASAR alternating polarization data,selecting the vegetation index DVI=0.45 as the threshold to divide the wetland of Poyang Lake into two parts: the vegetation with low leaf density and high leaf density.For DVI<0.45 the areas with low leaf density,there is high linear correlation between vegetation index and biomass,so we apply the statistical analysis to build a linear regression model using the samples.For DVI>0.45 the areas with high leaf density,due to the decline of the effects from the soil backscattering behind vegetation canopy,it can be used in C-band approximate microwave scattering models in the canopy to estimate these vegetation biomass such as Carex,Reed.Finally,the entire wetland biomass of Poyang Lake is approximately 2.1×10⁹kg.The accuracy of the result is higher than the previous result that only uses optical remote sensing data in high leaf density.Hence,combining the merits of optical and radar remote sensing can effectively enhance inversion accuracy of the entire wetland vegetation biomass.

Wetland ecosystem which has special hydrology,soil,vegetation and biology characters are places where land and water meet.Wetland is the landscape having highest biodiversity and important habitat which has highest productivity and many ecological functions.Tianjin's Binhai New Area(BNA) is located in the northeast of North China Plain,Haihe River downstream and at the center of the Bohai Bay.The Central Government of China and the State Council have made important strategic planning to push forward development of the Binhai New Area and make it become the Comprehensive Reform Experimental Zones.We can forecast that BNA will develop rapidly in future.With the growth in economy,the disturbance to natural ecosystem will enhance.Wetland ecosystem is the main natural ecosystem of BNA,however now the wetlands have been destroyed both in quantity and quality.The wetland ecosystem is facing serious crisis.Combining the integrated technology of landscape quantity analytical method with GIS technology,basing on TM images between 1979 and 2008,the FRAGSTATS software and the investigation data,dynamic changes in wetland landscape pattern of Binhai New Area are analyzed.The results showed that the total area of wetland landscape has not been greatly reduced.The total wetland area is 58.78% of Binhai New Area in 2008 and the landscape dominance is obvious.Much natural wetland converted into urban construction land.Littoral wetland decreased as construction between 2004 and 2008,more than 90% of natural beach was used up to May of 2009.Wetland patch number is increased from 137 to 704 meanwhile the average patch area decreased with the development of economy.Driving factor analyses showed that the main natural factor is the decrease of precipitation meanwhile increase of average air temperature.Main human disturbance factors includes aquiculture,city construction and coastal reclamation.

Net primary productivity(NPP),defined as the net flux of carbon from atmosphere into green plants per unit time,is a fundamental ecological variable to measure the energy input of biosphere and terrestrial carbon dioxide assimilation,to indicate the condition of land surface area and status of a wide range of ecological processes.In recent decades,model simulation becomes a major technique for NPP estimation over large areas,which can be divided into several categories as statistical model,parameter model and process-based model.As one of the essential process-based simulator models,the boreal ecosystem productivity simulator(BEPS) was adopted to simulate the forest ecosystem NPP in this paper.With support of MODIS images,meteorological data and so on,forest ecosystem NPP for Fujian Province of China in 2004 was simulated by BEPS model and validated by forest carbon sink investigation data.The simulated result shows that,in 2004,annual mean NPP of forest ecosystem is about 578.97g C/m²·a,and the sum NPP reaches 46.18×106t C;the order of woodlands NPP is: bamboo ≈ broad-leaved trees>fir>Chinese red pine,and the average NPP is 788.6gC/m²·a,780.0gC/m²·a,519.8gC/m²·a and 437.3gC/m²·a respectively.The time-serials analysis of monthly NPP demonstrates that the forest NPP declined a lot during the period from June to August,which is the highest NPP under usual conditions.The precipitation decrease from June to August in 2004 is most likely to be one of the major factors for NPP decline,after the statistic analysis of meteorological data with monthly NPP.On the other hand,the positive correlation between annual forest NPP and elevation is outstanding,where the determination coefficient reaches 0.99.This special phenomenon implies,to some extent,that with the increase of elevation,the interference from human activity is lightened and the forest ecosystem NPP is improved and sustained.In other words,diminishing human interference to forest ecosystem is one of the key measures to maintain the robust productivity of forest,further to mitigate global climate warming.Finally,the uncertainty of BEPS simulation was discussed in brief with the data and the simulator itself.

Large-scale Man-made Linear Features(LMLFs),resulting from nature remarking,have played an essential role in physical distribution and energy transportation.Though with a wide extension,those linear features primarily concentrate upon infrastructure field,including dams,bridges,highways,railways,metro-lines and pipes.As the development of spaceborne SAR technology,multi-baseline Differential SAR Interferometry(DInSAR) overcomes the spatial/temporal decorrelation and mitigates atmospheric effects in DInSAR procedures,revealing competent for large-area deformation measurement with millimetric-centimetric accuracy.This technology is capable of achieving temporal evolutions,which is significant for the stability management of observed features.In this paper,the LMLF's deformation phenomena,causes and derivative hazards are firstly described.After that,the multi-baseline DInSAR methods are applied for their ground deformation monitoring.Multi-source SAR data,including ENVISAT ASAR(Guangzhou),PALSAR(Lantau Island,Hong Kong),TerraSAR-X(Shenzhen),are used for experiments,the capabilities of the Permanent Scatterers(PS) method and Coherent Targets(CT) method are analyzed.The results demonstrate that,with different interferometric image pair combinations,the PS method is more suitable for the large number of SAR images because of the applied single reference image strategy,and vice versa.Generally,the wave penetrability and critical normal baseline increase with the wavelength.Therefore,when low coherent regions are encountered,the longer wavelength SAR data,such as L-band ALOS PALSAR,are preferred to generate high quality inteferograms,bringing about robust ground deformation results;in contrast,the short wavelength data are selected for high coherent areas,such as TerraSAR-X and ENVISAT-ASAR,to retrieve deformations with high accuracy.Taking the geometric and scattering characteristics of LMLFs into consideration,four viewpoints including thematic layers and GPS observation data,SAR images selection,PS candidates extraction and model improvement are discussed for deformation monitoring using multi-baseline DInSAR methods.Finally,from the experiments and analysis,one can conclude future works should be carried out especially in LMLF-oriented DInSAR model development.

China South-western karst rocky desertification has been more and more concerned due to its increasingly far-ranging influence on economical society and ecological environment.By remote sensing,a wide range of rocky desertification information can be practicably obtained.However,information extraction of rocky desertification based on remote sensing is not just an isolated technology,but a process of comprehensive analysis and feedback.To further explore the technology on information extraction of rocky desertification,this paper is mainly related to illustrating current developing of information extraction technology on rocky desertification,and on the basis of this,a new strategy for information extraction of rocky desertification has been raised.Specifically,we firstly analyze current researches of rocky desertification classification system,and then make a comparison between karst area and rocky desertification region.Furthermore,we illustrate the relationship between land use and rocky desertification,and summarize the current key technology of extracting rocky desertification information.Finally,we have drawn four aspects of conclusions as following: Firstly,due to huge differences in the aspects of classification amount,classification index and classification regions in current classification system of rocky desertification,it is useful to introduce the scale division method into the system,which can unify different classification types of rocky desertification levels.Secondly,the content-based characteristic discerned algorithm is a helpful research taste for karst area extracting,which can raise the precision of extracting rocky desertification information at last.Thirdly,it is an important way to transfer land use vector data into rocky desertification information by rationally using the corresponding relationship between different types of land use data and different rocky desertification levels.Fourthly,traditional research methods that building rocky desertification symbol,improving vegetation index model or establishing spectrum relationship model to extract rocky desertification information cannot well compromise the comprehensive characteristics of rocky desertification,so an extraction method based on rocky desertification comprehensive index has been put forward as a result.

Baiyangdian is the largest natural lake and the typical inland wetlands in the North China Plain.In recent 40 years,Baiyangdian wetland is facing serious environmental problems,such as water shrinking,ecological function degradation,biodiversity reduction,and water pollution.In this paper,by using the CORONA spy satellite image in 1964,Landsat MSS images in 1974 and 1983,ETM+ image in 2002,the land cover change and driving forces of Baiyangdian was analyzed.The eCognition 5.0 was used to classify the remote sensing images of the study area,and the spatial analysis superimposed tools in ArcGIS 9.0 was used to obtain the land cover conversion data in different years.The result indicated that the area of the Baiyangdian wetland is decreasing in the study period.The area in 1964 is 407.3km²,and 274.63km² in 2002.Furthermore,great change happened in the water area.In 1964,the water area is larger,with the value of 346.75km².In 1974,the water area decreased to 94.65km2.In 1983 and 2002,the water area is smaller,with the value of 67.27 and 46.86km² respectively.The main factor that causes the land cover change is the variation of water supply.There are two main reasons to impact the water supply of the study area.One is climate change,including the decreased precipitation and the increased evaporation.The other one is human activities,such as water consumption for urban development,land use change,the construction of reservoirs and water transfer projects.This paper applied the early CORONA spy satellite images to Baiyangdian wetland land-cover change analysis,solved the problem of lacking in data of early time period in the region.It made the research time back up to 1960s.Furthermore,the multi-temporal satellite remote sensing data used in the study improved the temporal resolution of land cover monitoring,and provided basic data for performance of regional land-cover change analysis.

During past five decades,land use and climate changed substantially in upstream basin of the Naoli River,Sanjiang Plain,resulting in huge hydrological changes accordingly.However,quantitatively distinguishing the hydrological responses to each factor-land use change and climate change,is not an easy issue.In this paper,the impacts of land use and climate change to hydrology in this region were assessed.First,based on the hypothesis of one factor changes slightly,we built the functions that hydrological processes responses to land use and climate change separately and forecasted the hydrological change resulting from cropland change and climate change respectively.Second,the impact of each factor to hydrologic change was quantitatively distinguished and then the hydrological processes driving force model was built based on precipitation and cropland area through least square arithmetic.Last,we gave a forecast model of hydrological processes with both natural and anthropogenic factors.The results showed that between 1956 and 1975,climate had a strong effect on hydrological processes.However,after 1975 the influence decreased gradually.Taking one with another,climate model behaved better to simulate annual average discharge than to simulate the peak discharge.Furthermore,cropland expansion that has being done since 1954 did not bring on significant change to hydrological processes.But comparing with annual average discharge,land use change impacted peak discharge more.Hydrological processes driving force model in this study was fairly ideal and it gave the modeling errors of 0.5 and 1.04,indicating the precision improved highly,which,was in contrast to a single factor model.In addition,the dynamic model performed more suitable for annual average discharge whose precision improved more with the R2 of 0.933.As a whole,hydrologic processes in the basin were mainly affected by climate change,while the effect of land use change was weaker relatively but unneglectable,especially,effect to peak discharge is increasing.

The article presented a Nearest Neighbor Index Fuzzy Clustering (NNI-FC) algorithm for ecological regionalization based on multi-layer grid model with consideration of spatial distribution of geographical factors such as climate, vegetation and topography. It's a "bottom-up" regionalization approach and solved the problem of how to determine the ecological regionalizations type and its boundary by calculating the Nearest Neighbor Index (NNI) and the similarity between grids. Numeric and non-numeric features were considered simultaneously in the NNI and similarity, so the algorithm integrated both qualitative and quantitative regionalization methods. The algorithm consisted of three consequence steps: data preprocessing, core region generation and fragmented region elimination. In the data preprocess, some numeric property values were transformed to non-numeric ones through classification of the combination of geographical factors. Next, core regions and fragmented regions were generated by ROCK algorithm, which clustering the adjacent discrete grids with the same property values. Then, on the basis of analyzing the fragmented regions area coverage and its spatial distribution by using NNI, the algorithm divided the fragmented regions into small pieces and merged them into the core region which has the biggest similarity. Finally, an eco-regionalization scheme for the given natural section is formed. The experiment of ecological regionalization for North of Xinjiang shows that the algorithm achieves over 80% classification accuracy and can be very good at expressing the diversity of regional characteristics. Besides, different levels of eco-regionalization scheme can be obtained by adjusting the thresholds of the algorithm and its time complexity is between linear and quadratic ones depending on the thresholds.

Sanjiangyuan Region, as the headwaters of the Yellow River, Yangtze River and Lancang River, is known as China's Water Tower. In recent several decades, continuously grassland ecosystem degradation and soil erosion in the region were increasingly serious due to natural and human activities such as global warming, overgrazing, mining etc. Currently, the degradation of grassland ecosystem has attracted attention worldwide. Therefore, the spatial and temporal patterns of soil erosion in the region were beneficial to provide scientific foundation for ecological protection and construction. This study aims to analysis the relation between grassland degradation and soil erosion, and spatial variations and dynamic of grassland soil erosion in the region since 1970s, according to remote sensing interpretation of soil erosion, land use and grassland degradation information. This paper analyzed spatial difference of soil erosion and its dynamic status over the past 30 years in grassland ecosystem in Sanjiangyuan Region, Qinghai Province. The results showed that the proportion of soil erosion in this region account for 46.74%, and 56.04% in the grassland. The primary types of soil erosion in Sanjiangyuan Region are composition of freeze-thaw and wind or freeze-thaw and water erosion, about 41.93% and 20.48% of total erosion area respectively, and most of the erosion degree is slight. The distribution altitude range were 3200~4600m, 2800~3600m and more than 4400m, and the slope range were 5~25°, less than 3° and 5~15° for water erosion, wind erosion and composition freeze-thaw erosion respectively. The erosion degree of wind erosion and composition freeze-thaw erosion increased with elevation rising, and decreased for water erosion. Soil erosion from the beginning of 1990s to 2004 showed more enormous degradation than the former period, and the warming climate and grassland degradation resulted overloading and overgrazing were the main driving forces.

Digital Elevation Model (DEM) is the main data source for digital drainage network extraction. But the extracted result would have great subjectivity because of the scale influences by the parameters such as the drainage area threshold and DEM resolution. Therefore, it is of great significance that how to optimize these factors for extracting the drainage networks better. The existing researches are mainly concerned on any two factors qualitatively or quantitatively. It is yet to be answered about how to define the drainage area threshold and the spatial resolution in a given research area. Taking Xiamen City of Fujian Province as the study area and using the Digital Line Graphic (DLG) on the scale of 1:10 000 as the original data to construct the DEM with grid cell size of 2.5m, we re-sampled the DEM into a series of DEMs with nested grid size as 5m, 10m, 20m, 40m, 80m, 160m and 320m using the Nearest Re-sampling method. And then the D8 algorithm and the Hydrology module in ArcGIS software was employed to extract the drainage networks from each DEM mentioned above with 7 drainage area thresholds as 0.9km², 2.7km², 4.5km², 7.2km², 9km², 13.5km² and 18km². Finally, the length of the drainage networks, as a key factor, was calculated to analyse quantitatively the relationship between the three parameters mentioned above and to further discuss the optimization of these parameters. The research results revealed that:(1)The length of the drainage networks obeys one composite power function;(2)When thresholds keeps constant, drainage network features decreased with DEM grid cell size continuously;(3)According to the mathematical analysis of the composite function, the optimal grid cell size and drainage area threshold are 40m and 7.2km² respectively. The extracted drainage networks with the optimal parameters met with those extracted from the SPOT5 remote sensing image very well.

Model coupling has been a prevalent research method in geoscience recently. Modeling framework is of the same category among various modeling environments by coupling, which gains many applications at present time. But the requirement of the frameworks decreases their usability, and also throws a heavy burden to users, which is researchers have to perform the coding and compiling tasks for coupling models. Taking ESMF as the foundation that is a representative modeling framework, this paper gives a detailed discussion on the problems presented in the research of building an icon-based modeling environment by coupling and their solutions. A methodology of generating codes of coupled models automatically is first proposed according to the coupling modes refined, as well as the relationships implicated by them which are the orders of calling models and the dependency of the input and output variables of coupled models. Then an approach of Open Concept Framework is developed with the purpose of expressing and confirming the semantic similarity between model variables. Based on the OCF and its building foundations, model metadata, a method is designed to verify the spatial and temporal consistency between the coupling models, and the matching between the input and output variables. Lastly, a prototype of icon-based modeling environment by coupling is developed, which implements the methodologies given above. Furthermore, two testing scenes are devised to examine the characters of the prototype. The results indicate that the entire error coupling operations existing in the scenes are refused by the system, and oppositely the correct couplings are accepted completely. Also, valid codes for the accepted coupling models are generated by the system and run in ESMF successfully.

The remotely sensed imagery classification is an important preparation work for geo-science analysis with remotely sensed imagery.The accuracy of classification will influence the quality of the result of geo-science analysis.We can hardly achieve satisfactory classification result using traditional classification method in high-resolution remotely sensed imagery classification.The research shows that the texture is a very significant feature for improving the accuracy of classification of high-resolution remotely sensed imagery.The time-frequency method of wavelet transform has the spatial advantage in texture analysis of remotely sensed imagery.In this paper,the texture feature of remotely sensed imagery is extracted from QuickBird imagery via wavelet decomposing.The feature space is composed of texture feature and spectrum response value of every band of the image.Then the image is classified by fuzzy C mean-value clustering(FCM clustering) approach.The experimental result indicates that the accuracy of classification can be increased when the texture feature is added into feature space.Meanwhile,the inner-homogeneity of a single category is improved.We can also see that the wavelet analysis can extract subtle texture better than coarse texture.In this paper,it is from the gradient image that the texture feature is extracted,so there are some misclassified pixels at the edge of the image,like the sides of roads.Some tiny objects can be amplified.The change of parameters used in this method can impact on the result of classification,such as the window size of the wavelet mode variance,the window size of mean filter,the coefficients of spectral responses,and so on.How to find out the influence mechanism of these parameters and how to choose proper parameters are what need to be resolved in next researches.

Urban heat island has close relationship with urban environment and global climate change.Using meteorological observation data to study urban heat island is usually to calculate the difference of the mean temperature between rural meteorological stations and urban ones.Therefore,it is important to distinguish the urban stations from the rural ones correctly,which directly affects the accuracy of urban heat island research.Previous classification methods,usually based on population census data of administrative regions,did not consider the spatial distribution of population.The city's large population cannot reflect the situation around the station,which may lead to wrong results.In addition,it is subjective to designate rural and urban stations by manual interpretation based on remote sensing and land use,which cannot reflect the combined action of land use to meteorological stations.In this paper,BP artificial neural network was employed to build a nonlinear model between the land-use types within 5 km around the meteorological stations by remote sensing and their rural/urban property.And the model was tested by population grid data obtained from remote sensing and statistical data.The simulated result fitted the stations' rural/urban property well.This model avoids the limitation from administrative regions' population census data and reflects the combined action of land use to meteorological stations.Furthermore,this model can combine the strengths both of the density of population and land-use types by a further analysis.In order to test the temperature difference between city and rural stations,we make a temperature comparison between some city stations and the surrounding rural stations.The difference is obvious.At the same time,we build the background temperature of Anhui Province in 2000 by rural stations classified by the model.Compared with the background temperature,city temperature is averagely approximately 0.4℃ higher.This model can be adopted in lager areas and even for the whole country,where has the characters of larger coverage and shorter time of remote sensing images.

The cloud had a significant influence to the quality of optical remote sensing images and the retrieval accuracy of land surface parameters.As one of the variable spatial-temporal factors,cloud would limit the application of optical remote sensed images.For the MODIS data with its swaths cover about 2330 km,the current metadata standards can only describe cloud quantity of the whole image,which limited regional studies and the application of MODIS images.Based on studies of current remote sensing metadata standards,we introduced a new metadata item named "regional cloud-cover" in order to describe the images' cloud spatial distribution,and extract the cloud cover information of specific areas from the MODIS level two cloud mask production(MOD35).This algorithm can extract provincial cloud quantity fast and accurately,and users can upload the vector boundary data or use tools provided by the system to draw their study area on the map,in order to specify the region that they study in the remotely sensed images,which would bring convenience to the regional research using MODIS data.In this paper,we realized the algorithm in the Visual C++ environment.First,we overlaid the girded vector boundary onto the raster data which has no projection.Then extract the cloud-cover percentage from sub-image that tripped from the original image.Finally we made a comparison between the results retrieved from the remote sensing images before and after geometric correction.It showed that,the geometric correction can improve the accuracy of the result little,but need to take much more time.We also made a comparison of the cloud cover percentage between the whole image and the regional one(Anhui Province),the result shows that the regional cloud-cover can represent the cloud-cover status of the specify region much more accurately than before.

As an important indicator of vegetation,vegetation coverage,which is generally estimated by visual measurement,is always used as an important index in vegetation evaluation and a key parameter for remote sensing inversion.But this method strongly depends on individual variables,and without the reproducibility of the results.That is to say,different observers almost certainly record different measurements with the same quantity.The advent of digital photography and automated image processing promises a revolution in the way vegetation coverage is measured.Recently,an automatically extracting algorithm for vegetation coverage was studied based on color features and some threshold values,and have a high accuracy when it is used to calculate the winter vegetation coverage with vertical digital photographs.However,this automatic method isn't suitable for Tibetan Plateau due to its various types of vegetation and background.Though analyzing on the color characters of vertical photographs,we found a named excess green method,which is sensitive to green vegetation and enhances the contrast between plant and soil background.A k-means clustering algorithm was designed to divide photographs into plant and background,and calculate vegetation coverage automatically.Compared with the result from the manually supervised classification method,the root mean square error was less than 5%,but it spend less time than supervised classification and had a higher accuracy than that of visual methods.Moreover,some approaches to improve classification accuracy were discussed by analyzing the error source of automatic classification.

In order to integrate distributed and heterogeneous earth system scientific data resources and provide one-stop data sharing services for different users,a distributed data sharing platform is needed urgently in China. In this study,we analyzed interoperability requirements of the distributed earth system scientific data sharing plat- form(ESSDSP)in the first place.After compared with several architectures and technologies of distributed inter- operability,we chose SOA(service-oriented architecture)as the architecture of ESSDSP(hereafter called SOA4GS,SOA for Geo-data sharing).The basic principle of SOA4GS was to abstract a series of web services from the whole data sharing activities which included information synchronization service of data centers,distributed metadata synchronization management service and multi-resource data access service.Through deploying these web services in different data sharing nodes,the software can easily implement the interoperability among them.Finally we introduced the design,implementation and deployment of these web services in detail.SOA4GS has the following characteristics:(1)Distributed nodes with uniform logic ESSDSP are comprised of many centers which are deployed in different organizations at different locations.They have the uniform business logic to ensure they can communicate each other by web services.(2)With open and extensible architecture based on SOA,ESSDSP is an incompact architecture.So other systems based on web can be easily integrated into the platform after little function model revision for compliance with the interface of web services,while,on the other hand,systems such as e-government and e-commerce can use data sharing functions by calling web services to share the data provided by the platform.(3)Easy to use Web service platform is a web software module.Its client code can be created automatically from the web service' address by using some development tools such as Eclipse,JBuilder.So users can easily call the platform's web services or revise their system's functions.In practice,ESSDSP has been used in Earth System Scientific Data Sharing Network,one of National Infrastructure and Facility Development Environment Building for Science and Technology Industries Program of China.It comprises one main center,one authentication center and sub-centers.So far,it has integrated 10.23 T Byte data resources and Byte data resources have been downloaded by 34,597 users.

In this paper,a method of ecological security evaluation in Xinfengjiang River Valley was studied,and a theoretic framework was set up with RS and GIS technology.Based on ALOS image,10m×10m grid was taken as basic unit,while township as comprehensive unit.Then Pressure-State-Response(PSR) framework was improved and GIS was used to acquire 18 evaluation factors information,such as settlement pressure,road pressure,mining pressure,water quality,vegetative cover index,rivers density index,land deterioration index and so on.Ecological security index was calculated next by different models in every grid and every town in order to make ecological security classification maps.According to the study,ecological security index is divided into six levels.levelⅠmeans very safe,and level Ⅵ means very unsafe.From levelⅠto level Ⅵ,the ecosystem services become less and less perfect,the ecosystem structure become less and less complete,and the capacity of ecological restoration and reconstruction reduce gradually,while the destruction degree of ecological environment and the possibility of ecological disaster increase gradually.The results show that(1) Ecological security in Xinfengjiang River Valley is good from global aspect.It accurately reflects the status of ecological security in Xinfengjiang River Valley.(2) Security level distribution in was not balanced,which has evident regional characteristics.The worried status of ecological security were in residential、industrial and mining area,while states of Xinfengjiang reservoir area were better.(3) 82.43% of this area is in safe,while about 17.57% of the area is in warning.Despite the overall relatively good situation,insecurity has also reached an area of about 217.600 8km2,which accounted for 3.8% of the watershed area.Of the 41 towns in the river valley,there are 35 towns are safe,5 towns are in warning,with only one town unsafe.The study also shows that RS and GIS technology has positive significance for protecting and improving the ecological security in Xinfengjiang River Valley.

Soil is the most basic and most important resource for the living and development of human beings.At present,there is an urgent need to obtain accurate soil spatial information for environmental modeling and agricultural applications.The theory of soil-landscape relationships relates difficult-to-measure soil information,which includes soil type and properties,with some easy-to-obtain soil-forming environmental factors.This makes it possible to infer soil spatial variations from the easy-to-obtain environmental factors.Satellite remote sensing techniques generally serve as auxiliary tools for researches on soil spatial information extraction.The techniques can be used to acquire landform and vegetation data.Then,based on their relationships with soil conditions,soil spatial variation can be derived.In areas with median to strong landform gradient,easy-to-measure landform and vegetation factors have played an important role in soil mapping.However,for the plains and topographically gently undulating areas,where spatial variation of soil properties such as soil texture usually cannot be effectively indicated by the landform and vegetation,a scientific problem is how to reveal spatial variation of soil properties.This paper presents an approach to using land surface dynamic feedback captured by high temporal resolution remote sensing over a short period immediately after a major rain to distinguish soil texture over such areas.The results of this study showed that land surface dynamic feedbacks captured by high temporal resolution remote sensing can be used to identify spatial variation of soil texture.Locations experiencing different soil texture have obviously different dynamic feedbacks,and the more similar the soil texture between locations the more similar their dynamic feedbacks.The findings are very encouraging.The work reported in this paper provides an effective solution to the difficulty of identifying soil spatial variation over low relief areas.It exhibits the potential of high temporal remote sensing on the exaction of soil information.

LUCC informatics system,LUCCIS,is a newly developed research field on the basis of integrating the classical methodology of physical geography and the new progresses in GIS technologies.As a research field focused on studies of temporal and spatial processes of LUCC,LUCCIS has developed rapidly in past decades in which some technical components have been improved greatly.LUCCIS focuses on material cycles and macro life processes of multiple scales on the land surface with supports from remote sensing,GIS,and data fusion technologies.In this paper,the progresses of those kernel methodologies,techniques as well as compositional research achievements in China are described in detail,which make it possible for the establishment of LUCCIS.Specifically,the development in the technologies of extraction and analyses of remote sensing information,the exploration of driving mechanism of LUCC over space and time dimensions,the simulation of the spatial-temporal process and pattern of LUCC,the evaluation of the macro-ecological effects of LUCC and the establishment of research information platform of LUCCIS are all promoting the development of LUCC studies in China.The domestic research trends indicates that the development of extraction and analytical technologies of remote sensing information,the case study of exploring driving mechanism of LUCC,the methodological innovation of simulating the temporal and spatial process of LUCC,and the more accurate description of macro-ecological effects of LUCC play important roles in the methodology framework of LUCCIS.Also,the construction of multi-source and temporal-spatial data platform and the realization of massive data availability and integration are all kernel components of the methodology framework of LUCCIS.We can expect that LUCCIS would play a more and more important role in promoting LUCC research activities at regional and global scales.

2009年11月25日是我国著名科学家陈述彭院士逝世一周年的纪念日。一年来,陈述彭院士在各种场景的身影,同我亲切交谈时的面容,特别是在医院的最后几次见面交谈时的难忘情景,时常出现在我的脑海,悲痛、回忆与思念经常交织一起。我自1961年7月从苏联莫斯科大学地理系地图学专业毕业回国,被分配到科学院地理研究所地图研究室工作,时任地图研究室主任的陈述彭先生对我非常信任和器重,当即委任我担任专门地图学科组副组长,并负责国家自然地图集的统一协调与土壤图组的编辑,1962年还安排我在全国地图学术会议上作大会学术报告。

The process of spatial data mining is closely related to the process of spatial concept formation and analysis. From the view of recognition,centered in the thought to extract the concepts in different hierarchies and granularities, based on the formal theories of concept formation and analysis,i.e.concept lattice and cloud model, this study analyzes and summarizes the methods and their application of spatial data mining based on concept analysis, including the methods and their application of spatial association rule mining based on concept lattice,those of conceptual clustering based on concept lattice,those of spatial association rules mining based on cloud model, those of spatial clustering/classification based on cloud model,etc.Concept lattice provides a kind of theory which can express and analyze spatial concepts and their concept structural relationships by mathematical and formal methods,which is good for the understanding of human,and is convenient to process by computers;Cloud model is a kind of uncertainty theory which can deal with fuzziness,randomness and their association,it is a more powerful and effective uncertainty theory than fuzzy sets and rough sets.The further research directions of spatial data mining based on concept analysis are as follows:(1)the formal expression of spatial concepts;(2)the formal analysis of the process of spatial concepts formation;(3)the formal expression and analysis of the structural relationships of spatial concepts;(4)the construction of the spatial concepts hierarchy and their formal expression and analysis;(5)the expression and analysis of spatial concepts and the concept structural relationships with uncertainty .

Geo-spatial relations reflect the complex association between geographical phenomena and environments,so,are very important in dealing with geographical issues.They can be generated according to geometric attributes of spatial phenomena,such as distance,direction,connectivity,similarity etc.Also they can be generated based on combinations of geometric attributes and non-geometric attributes such as statistical correlations,spatial autocorrelations,space interactions,spatial dependencies etc.There are still other geo-spatial relations arising from non-geometric attributes of spatial phenomena.Extracting geo-spatial relation rules effectively helps to improve the accuracy and efficiency of solutions of geographical issues.This paper discusses both the expression of geo-spatial relations and the extraction process of geo-spatial relationship rules based on rough set theory.To illustrate the extraction process of the rules,we take the land use vector data of the Pearl River Delta in 2000 as an example to extract the geo-spatial relationship rules of urban areas and rural residential areas in Shenzhen and Hong Kong.

The principal advantage of hyperspectral remote sensing is to identify composition and components in objects and inverse the physical and chemical parameters according to the acquired and rebuilt pixel spectral signatures.This paper reviewed the composition and components of rocks,soils,atmosphere,vegetation,water body and manmade structures that can be recognized with hyperspectral remote sensing,described spectral identification methods and then,discussed some key characteristic techniques comparing with multispectral remote sensing,such as spectral matching,spectral unmixing,endmember selection,and quantitative inversion as well.