Artificial vegetation is the important bio-component with absorbing dioxide carbon for supporting ecosystem health.Artificicial carbon sink forest play important function on reduction of carbon dioxide emission.Based on TM data of August,2009,the eco\|landscape pattern information is obstained.Furthermore,using NDVI to estimate vegetation carbon density,which is a viable method.Through investigation the biomass of arbor layer and herbage layer,the carbon density of arbor layer and herbage layer are determined.The result as follows,the average carbon density is 37.04 mg/hm2 in the arbor layer,meanwhile,the average carbon density is 59.65 g/m2 in the 1 m×1 m sample site in the herbage layer.So,the total vegetation carbon density above the ground is about 37.64 mg/hm2.As a result,the carbon reserves is 250 915.5 mg in the whole vegetation layer.With the vegetation growth and biomass accumulation effects,the carbon sink function of the artificial vegetation will be further more bringing into played.
The actual ET was simulated in this study to estimate and analyse the time\|space distribution of ET in TaiWan.The study improved the scheme of look\|up table of monthly roughness length,zero plane displacement and vegetation height related with land covers based on the surface energy balance system(SEBS),and set up the simple correlation between land surface temperature and air temperature in Taiwan.Then,a modified quantitative remote sensing ET model,SEBS\|China model,was developed based on SEBS model.The SEBS\|China model simulated the actual ET by using meteorological data of 8 selected dates in year 2002 and 2003 with MODIS data.Comparing to the Pan ET,the modeling values are slightly lower,which is consistent with the fact that Pan ET is potential ET,should be higher than actual ET.Based on the results,several analyses are made to deeply understand the characters of the ET in Taiwan area.Namely,there is an obvious tendency in the temporal and spatial distribution of Rn,EF and ET,namely,the values of the three parameters are higher in central mountain areas,and reduce gradually to the east and west.Through a year,the ET is the highest in July,and lowest in January.According to different land surface cover,the ET is different; generally,the land cover from high to low is forest,grass,water,bare land and urban area.Comparing to the rural areas,ET of cities is much lower.
Apparent thermal inertia model was used to retrieve the soil moisture on the basis of MODIS image data from April since 2009 to 2010.First of all,the relationship between the ATI and soil moisture data was fitted with a model,and then combined with the date of 37 measured soil moisture points,the distribution of soil moisture condition was obtained in Fuxin.Finally the soil moisture distribution of Fuxin area in the spring time was acquired with the linear model.The field data validations show that,at the depth of 0~10 cm,the mean relative error is 5.4%;at the depth of 10~20 cm,the relative error is 10.82%;at the depth of 20~30 cm,the relative error is 16.50%.Although the relative error can meet the needs of the actual inversion,the relative error increases as the soil depth increases.
With the development of interference imaging spectrometry,ground applications have posed higher requirements for higher processing rate and real time processing in spectrum reconstruction.In this paper,reconstruction methods and CUDA programming are especially discussed,a spectrum reconstruction technology based on the developing CUDA technology is presented.Experiments show that this method can significantly improve the processing rate of spectrum reconstruction and meet the demands of ground applications.
Most of the hyperspectral image enhancement methods,which are applied to multispectral image processing,have not fully used the spectral information of hyperspectral data.The unmixing based enhancement method is limited by endmember selecting.A fractal signature based approach to image enhancement is presented in this paper.According to modified blanket method,the upper and lower fractal signature curve corresponding to original spectral curve of each pixel can be computed,and the formed fractal signature image at each scale is used for image enhancement.Three scenes of Hyperion image are experimented in the work,the corresponding upper and lower fractal signature images are computed.It can be seen from the fractal signature curves and images that ground targets are characterized differently by signature value and feature scale.The experiment result shows that the fractal signature image can outstand targets more saliently than the original data.Thus the proposed approach can be applied to hyperspectral image enhancement.The experimental result shows that the original curve,the selection of initial scale and number of sample has a significant impact on the fractal signature,and a brief discussion on these issue is presented.
The atmospheric correction for hyperspectral remote sensing images is absolutely necessary for improving the precision of data pre\|processing. The emphasis and difficulty is how to select the appropriate method of atmospheric correction for hyperspectral remote sensing based on different data,different research areas,and different research purposes.In this paper,logarithm residual model,empirical linear regression model,FLAASH model and ACORN\|3 model are used as the atmospheric correction models.The Hyperion hyperspectral image is used as the original data to analyze the precision of atmospheric correction.In order to evaluate and validate the correction results,we use the object spectral response curves and spectral matching to analyze the original images and the images after atmospheric correction.As a result,the reflectance of ground object by the method in the paper can eliminate the atmospheric scatter influence effectively and the edges of the surface objects are distinct and easy to identify.Furthermore,the ACORN\|3 atmospheric correction model is better to Hyperion hyperspectral remote sensing images than the other model.
This paper explored the feasibility of developing an 8\|day time serial MODIS\|NDVI by merging the phased 8\|Day Terra\|MODIS\|NDVI (MOD\|NDVI) and Aqua\|MODIS\|NDVI (MYD\|NDVI) 16\|Day time serials in Collection 5 (C5).First,MOD\|NDVI and MYD\|NDVI time serials were filtered separately based on Quality Assessment (QA) and phenology,which ensured cleaned NDVI time serials.Second,the actual day of pixel in each composite period was calculated,and the results showed that the percentage of pixels in first 8\|Day of a 16\|Day composite period was 55%.The further analysis indicated that if MOD\|NDVI and MYD\|NDVI were simply merged,7% of pixels would suffer from time disorder.This time disorder might lead to NDVI disorder,which might not be reasonably restored by filtering.Finally,a conceptual framework was presented to double MODIS\|NDVI temporal resolution based on C5 Terra and Aqua.The priority to select a pixel for an 8\|Day composite period was actual Day of Year (DOY),QA,and NDVI,respectively.
Based on the features of the different mathematical models during the image processing,we proposed a remote sensing image fusion method,which is based on Curvelet transform and IHS transform.Firstly,the IHS transform is applied to multi\|spectral image; then,the I component and the high\|resolution image are fused through Curvelet transform; finally,we can obtain the result by inverse IHS transform.The experimental results show that the method has outstanding advantages,which maintain the spectral characteristics of the source image,and its results can be applied to the field of spectral analysis.
New snow cover product is produced by the combined using of MODIS Aqua and Terra snow cover products from 2001 to 2008.Then two sets of MODIS snow covered days are produced by the using of new snow cover products and Terra/MODIS snow cover products.Through comparison of this two different snow covered days in North Eastern and Inner Mongolia region,it indicates that the differences of snow covered days between these two sets of data are growing with the rising of elevation.The differences between these two sets of data are about 15 days in average.Although the combined use of Terra and Aqua products can generate more snow covered days,the combined data still lower than insitu observations of snow covered days about 27 days.Since the combined snow covered days smaller than insitu observations,but with a good regression relationship,which means calibration of MSCD image could significantly improve its precision status in those regions.After the calibration of MSCD based on insitu observations,the difference between insitu SCD and calibrated MSCD are reduced from 27 days to 18d.
Endmember extraction is an important and challenging task for hyperspectral image analysis,and it is also a critical step to resolve the problem of hyperspetral image unmixing.For the current endmember extraction algorithm,the abnormal pixeles were considered as endmember at the same time on the process of endmember extraction.How to distinguished the abnormal pixels and endmembers effectively was not only a bottleneck of hyperspectral endmember extraction,but also a key factor of improving the accuracy of hyperspectral unmixing.This paper presented a hyperspectral endmember extraction method that supported by anomaly detection.At first,RX algorithm was used to detect the abnormal pixels of original image,and determined the number of anomaly pixels by the detection image,then replaced that pixels value by the mean vector of original image.Secondly,Orthogonal Subspace Projection (OSP) was used to extract endmembers from the solved image.Experiments showed the method can distinguish the abnormal pixels and endmembers effectively,and inhibited the abnormal pixels adding to endmember extraction process.This method takes so little influence to endmember extraction and can remove the abnormal pixels.It proved that removed the abnormal pixels before endmember extraction was effective.
Firstly,wind direction from Seawinds scatterometer was used as the initial wind direction for the wind field retrieval from synthetic aperture radar images,which includes the process of noise removal,radiometric calibration,polarization conversion,acquisition of normalized radar backscatter coefficient and ocean wind field retrieval and display by using geophysical model function.Secondly,wind direction form numerical weather prediction (NWP) model WRF (Weather Research Forecast) was used as the initial wind direction to get the wind field from SAR images.Comparison of results form the two ways were made to validate our methods.Combination of high resolution NWP wind direction and SAR images will be the trends of operational coastal wind field retrieval in the future,a useful try was carried out in this paper.
Snow is one of the most active factors in the nature which influences the redistribution of solar radiation energy and water resource.The distribution of snow cover over mountainous area of Manasi river basin will exert great impact on river runoff in snowmelt season.Based on snow cover frequency and snow cover fraction acquired from remotely sensed MODIS/Terra snow cover product from 2000 to 2010,the study discussed the spatial and temporal distribution of snow cover in mountainous area of Manasi river basin.There are abundant snow cover accumulation in January with several high snow cover frequency areas appear in lower altitude area in the north,moderate altitude area in the south and the source of Qingshui river and Taxi river.In general,the snow cover distributions show the same pattern and the snow cover frequencies both rise with altitude in April and October,while only a few areas with high altitude are covered by snow accumulation in July.Additionally,the snow cover frequencies maintain at high level throughout the year in the birthplace of Manasi river,Jingou river,Qingshui river and Taxi river,while it keeps in lower level in valley areas of middle and upper reaches of Manasi river.Furthermore,the snow cover fraction in January decreases from 95% in areas below 1 400 m to 41% in 2 600 m,and hereafter increases to about 80% in areas above 5 000 m.The snow cover fraction of January,April and October show similar tendency in most elevation zone with the highest value in the north,northeast and northwest slopes,followed by the east and west slopes,and the lowest in the south,southeast and southwest slopes.
Based on the MODIS satellite image data in 2001 and 2009,dynamic change of desertification in loess plateau was studied by employing methods of normalized difference vegetation index and the dimidiate pixel model principle.It focused on analyzing of the spatial distribution of the different desertification degrees and change in the area during both the August 2001 and 2008 by establishing a scientific classification system of desertification.The results show that desertification land area in the loess plateau exhibits over all a decreasing tendency since recent eight years,but development intensity continuously increased.The very severe desertification land area increases by 16.53 km2 with the growth rate of 28.36%.The severe and moderate desertification area decreases by 1.2×104 km2 and 7.0×104 km2 with the change rate of 32.97%and 29.19% respectively.9.0×104 km2 moderate and 1.2×104 km2 potential desertification land are turned into other types of desertification,and increases by 9.3×103 km2 and 7.3×104 km2 with the growth rates of 4.2% and 57.3% respectively.The developing area is 1.9×105 km2,steady area is 4.0×105 km2 and the reverse area is 2.8×104 km2.The developing area is 1.6×105 km2 larger than the reverse area,indicates that the environmental quality is continuously decreased,and the desertification degree is still expanding．
Remote sensing technology can detect alteration information.Choose Danghe\|Nanshan district as the study area,based on ETM and ASTER remote sensing data,using principal component transform,mask analysis,threshold segmentation method,ectraction alteration information from the Landsat\|7 ETM and forecast possible mineralization area.The method is accurate,effective,eliminating cloud,snow and other interference information,the use of the information obtained with the known mineralization alteration minerals point is very consistent,the method is helpful for extraction of alteration information.
Fog is a type of weather phenomenon with great damage which has serious influence on human activities such as transportation and so on.It is of great significance for fog monitoring.Fog area can be monitored by satellite remote sensing technology,according to the spectral characteristics among fogs,clouds and underlying surface in visible,mid\|infrared and thermal\|infrared channels.The daytime fog experiment has been made by multi\|channel threshold method with the VIRR data of FY\|3A satellite in the east China on April 5,2010.And the characteristic parameters of fog,such as the optical thickness,the vertical thickness and the visibility have also been estimated in this paper.The result shows the feasibility of fog monitoring with FY\|3A satellite data,and that the method is suitable for promoting operation of fog detection.
Grasping the information of cultivated area in time is extremely important to agricultural production.Its extremely hard to monitor a wide range of cultivated land with high\|precision.To solve this problem, this paper provides the solution of monitoring a wide range of cultivated land with high\|precision,which is sampling the high\|resolution,mid\|resolution,and low\|resolution of remote sensing images step by step,and using farming division as homogeneous area.Take Shangdong province for example,we take the use of MODIS images,CBERS\|2B CCD and HR images to sample,calculate the accuracy among different images,and correct accuracy step by step.At last the cultivated area of〖JP3〗 Shangdong province is calculated, and the data is only 8.5% higher than statistical data in the same year.The result indicates that the method is in good operation,raise the accuracy,and also satisfies the demand of monitoring a wide range of cultivated land.Therefor,it can be well used in agriculture production and research.
Among the various categories of landform evolution,sand dune is a kind which moves faster than the other.Until now,it is still fewer reports on this aspect while the dynamics mechanisms of sand dunes have been well studied.Its difficult to study the large area movement of sand dunes by using traditional method such as ground\|based observations.Since Landsat satellites launch in the 1970s,it provides long\|period,large scale images for the research of dunes movement.In this article,many temporary and scenes of Landsat Thematic Mapper imagery from 1989 to 2010 are used to analyze the movement of sand dunes in four typical areas of Badain Jaran desert.Direction and speed of sand dunes can be extracted from sand ridges of images after geometric correction.We get following results:① Landsat TM images can be effectively used to the epochs of ten years dynamic detection of sand dunes in the Bardain Jaran desert;② The movement of sand dunes in North are not distinct;③ Sand dunes in East and Southeast sand belt move towards SE,while sand dunes in south sand belt reciprocate from NW to SE.Sand dunes movement is consist with regional wind fields;④ The relation between the speed and the height of sand dunes can be analyzed with DEM data.There is a negative correlation between them.
Fractal is an important method to improve effect about geochemistry and geophysics,but currently fractal theory is rarely used in the field of remote sensing.In this paper,the multi\|metal mining area in Lushi,Henan is selected as the studying area ,and the ETM+ is the data source.Using the method of MPH Masking,Principle Component Analysis(PCA),Hue and Saturation Transform (HIS) (MPH)) extracts alteration information from the studying area.The method includes three aspects:① mask,② PCA,③ IHS transform.Next,adopting the method of summation in the fractal theory:the pixel brightness\|area model make a quantitative analysis about the extracted iron alteration form MPH.And then compare the result with the common methods of threshold selection by standard deviation.It is proved in this paper that the alteration information as a geologic record of spatial distribution has not only fractal characteristics but also multi\|fractal characteristics.The research results show that fractal technique is an effective method to classify alteration information of remote sensing data,and it makes up for the weakness of threshold methods and achieves better results.
Remote sensing technology has been widely used in disaster reduction and response works in earthquakes,floods and typhoons.It plays an important role in stage of warning,monitoring,rescue,assessment and brings us a new way to know the disaster better.Combined with its own characteristics in accessing information,this paper analyzes the advantages of remote sensing technology in disaster reduction.With the detailed description of their data parameters,it introduces the common remote sensing data used in disaster information access and summarizes the application status of them.This paper also discusses the major role of remote sensing in different disaster management phases,such as warning,monitoring,relief,assessment and recovery.After analyzing the data characteristics and its processing methods,it introduces some practical application in different type of disaster.At the end,the application and development trend of remote sensing technology in disaster reduction and response has been discussed.
Ground object spectrum is one of the most important remote sensing information sources.Recently,building a spectral database to manage these spectra data has attracted more and more attention.The database of natural resources and geospatial basic information is one of the four basic information databases of national E-government project.The main purpose of this database is to achieve the data sharing and exchange among different departments.National typical ground object spectral library is included in this database.To meet the requirement of E-government,we built an E-government oriented national typical ground object spectral library service platform.This platform aims to provide E-government suitable data and strategies.This article presented the design of spectral library service platform and related critical techniques.
Topological relationship is one of the most basic and the most important relations between two spatial objects in GIS.However,current topological models have deficiencies when they are used to describe detailed topological relations.In this paper,one general descriptive form of topological relation model was proposed based on the node degree and the euler-number aiming at random two objects.In this model,all parts of one spatial object were treated as a whole object,and intersection& difference operators were selected from several set operators to distinguish the topological relations,and four types of topological invariants were used for the computational results of contents,dimension,node degree,and euler\|number.The topological description form of line\|line,line-area,area-area were discussed emphatically.
The study was based on the development platform of Visual Basic 6.0.The secondary development of GIS was progressed by ArcEngine 9.3 component library of ESRI,analyzing the actual demand of geological hazards information management in China coastal zone.The information management system of geological hazards in China coastal zone was established.The system had the functions including the query and management of spatial data and attribute data,space statistical analysis,thematic maps drawing and hazard assessment,data conversion and output,etc.It provided the integrated services for geologic hazard prediction,forecast and prevention.
The main part of Tonglv canal in Nantong is selected as a sample.The information system for the management of pollution sources is designed by using the technology of component development,basing on Visual Basic6.0 and Map Objects2.4.It combining GIS technique and water environment model,realizes the functions of managing the data of pollution sources,calculating the model of pollution load,monitoring statistics and monitoring the pollution sources on the spot,and visualizing one\|dimensional simulation,etc.It helps decision\|makers with urban environmental pollution index,propose restrictions on the basis of emissions sources.In a certain extent,the system makes the monitoring and management of pollution sources more efficient and thus promote the information construction of water environment of Tonglv canal.
