After the 7.0 magnitude Sichuan Lushan earthquake occurred on Apr.20th,2013,the secondary geological disasters caused many damages of highway.It was an important and basic job of emergency relief for highway disaster monitoring fastly and accurately.There was a mature and effective way of obtaining highway disaster information based on remote sensing images and information extraction technology.Utilizing domestic remote sensing satellite images as well as airborne aerial images acquired by Institute of Remote Sensing and Digital Earth,Chinese Academy of Sciences,it focused on highway disasters monitoring with image\|based visual interpretation and analysis methods.After information extraction,433 geological disasters were found around roads and collapse was the main types of disaster.197 disasters have impacted highway and the total length of damaged sections was about 10 km.Meantime,there were 63 disasters became a serious threat to highway,which should be paid attention to analyse.The field investigation proved the road disaster extraction results were reliable.The highway disaster monitoring result was provided to the Ministry of Transport of People’s Republic of China within two days after the earthquake,and it was a sufficient basis for carrying out road repair scientifically and effectively.
ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) has been used extensively in alteration mineral mapping.However,using multi-methods and comparing the results to select the optimal areas,have rarely been reported.Sandaoling in Balikun County,Xinjiang was taken as the study area.Both band ratio and principal component analysis were adopted to map alteration minerals using ASTER data.Multi-information such as ore points,small intrusions,structure,geochemical anomalies were then applied to compare with the remote sensing anomalies to obtain the best results.Fifteen ore exploration targets were located by synthetic analysis of multi\|information including remote sensing anomalies,geochemical anomalies and geological data.Through ground inspection of nine targets,seven new mineralization occurrences were discovered.The result shows that using multi\|methods for alteration mineral mapping and comparing the results to select the optimal ones are more effective than using a single image processing method.
The conventional mathematical morphology method using single data source to extract road network which could not take full advantage of the road characteristics,the extracted road information was not complete.In view of this drawback,and base on the airborne LiDAR data can provide elevation information,this paper proposes a method which combines the airborne LiDAR data with high resolution remote sensing images to extract city road network.The airborne LiDAR data and high resolution remote sensing QuickBird images of Xuzhou were taken as the experimental data,the precise registration between them were first done,then the FRIR (False Road Information Removing) method was used to remove the vegetation and buildings separately,so the basic road contour was displayed.Finally,this paper achieved an Improved Road Trimming (IRT) algorithm under the ArcGIS and Matlab programming environment,the road network was trimmed by the algorithm,then a smooth and continuous city road network was obtained.The result of the accuracy evaluation indicates that the method was proposed can be used to avoid the influence of the building shadow,city squares,parking lots and the vegetation groups on both sides of the road to the road centerlines extraction well,and the recognition accuracy of the road network is more than 84%.
It is significant to monitor cyanobacteria bloom accurately and rapidly with remote sensing methods by multi-sources satellite data, In this paper,it took NOAA/AVHRR data which has lower spectral resolution of 1 100 m and Terra(Aqua)/MODIS data which has higher spectral resolution of 500 m as the main data sources.A region-based wavelet transform fusion method (REFS_wt) was presented which took the average as the low frequency coefficient and the high frequency coefficient was calculated by regional energy method.Then a comparison was made between the presented method and another image fusion scheme based on regional pixel energy(REFS_pl).The experimental results showed that the REFS_wt method was better than the region-based pixel energy fusion method in improving spectrum quality of remote sensing images.This studied method has been effectively applied to cyanobacteria bloom monitoring and early-warning for Lake Taihu.And the fused images could preserve all useful information from primitive images and that the clarity and information quantity were improved.
Fly ash not only pollutes environment,but also endangers human health.It is of great significance to the protection of the environment and human health that distinguish fly ash rapidly,real-timely,accurately based on remote sensing.The paper chooses Landsat 5 TM image as the source data and analysed the spectral information of the typical surface features in Baotou City. Then the study adopts the decision tree classification to extract fly ash in the study area.Firstly,analyzed the spectral characteristics of the typical objects and compared the relationship between them in the study area.Secondly,classified the image respectively through established the decision tree,used Soil-adjusted Vegetation Index(SAVI),Modified Normalized Difference Water Index(MNDWI),Normalized Difference Built-up Index(NDBI) and Spectrum Threshold Method.Ultimately,conducted post-process for the result of classification by shape feature and space feature.The total classification accuracy was up to 70.7%.The experimental results show that the method is suitable for the automatic extraction of fly ash information,can achieve high accuracy combined with the visual interpretation.
The assessment for earthquake disaster has important practical significance to the relief work,the traditional methods are mostly on the field statistic data after the earthquake,the shortness including the up to date state of the data acquiring and usually exists the efficiency of the disaster assessment.On account of this callback,and combined with the characteristics of the remote sensing data,this paper proposed an assessment method for post-earthquake urban damaged area based on the impervious surface estimate.Firstly,the two middle-resolution remote sensing images before and after the earthquake in the area of Yushu were preprocessed,and then a linear spectral mixture analysis method was used based on the VIS model,the endmembers of the remote sensing images before and after earthquake were extracted by respectively,and the abundance map of the impervious surface was obtained,finally,the earthquake disaster damage area assessment was evaluated by calculating the number of pixels of the impervious surface before and after the earthquake.The accuracy analysis indicated that the method was proposed this paper could obtain the macro situations of the damage area,includeing the urban buildings,important roads,etc.in quickly and timely,and provided a certain basis for post-earthquake reconstruction and decision.
To make up for the missing information over cloudy area of satellite remote sensing dust product,while improving its precision,visibility data monitoring by routine meteorological stations is fused with Infrared Difference Dust Index (IDDI) of FY-2 meteorological satellite.At first,the linear correlation between visibility and IDDI is built by data from each month of March,April and May from 2006 to 2010.Then visibility is transferred to IDDI with the developed correlation.And finally the Inverse Distance Weighted (IDW) method is applied to make them fusion.To pixels with IDDI values,visibility is used to improve its precision while to pixels without IDDI values,visibility is integrated to complement the missing information.This method is applied to dust process which happened in west-northem of China on the May 10,2007 and April 20,2008 and well demonstrates its accuracy.The comparison also showed that the process is more suitable for greater intensity dust,especially in the case that surface meteorological stations are located near the edge of satellite observations.
After Very High Resolution(VHR) radar satellite missions like TerraSAR-X,Cosmo-Skymed and RadarSat-2 have been launched in 2007,spaceborne Synthetic Aperture Radar(SAR) images reach spatial resolution below 1 m in azimuth and range in spotlight mode.In such kind of data,radar signals representing single building are disturbuted over a high number of resolution cells,which makes SAR to monitor urban areas possible.However,due to the complexity of urban environment,interpretation and application of SAR data still turns out to be challenging and time consuming.SAR images simulation,which can reduce the difficulty,is becoming increasingly popular.A novel method of SAR images simultion based on extended ray tracing procedure was introduced.The simulator,which has an important role in analysis of images,can effectly simulate scatterring effects of SAR images,such as layover,shadow and multiple bounce.Firstly,a virtual SAR system was formed.For the sensor,rays were transmited and received.The pathes of the rays and the interaction with objects were traced by the simulator.Signals from different reflection levels could be stored in seperated image layers.For evaluating the effectiveness of the simulation approach,simulatd images were compared with real TerraSAR-X images for several models like a building with plane-roof and gable-roof,National Gymnasicm and an extended scene containing numbers of building models.The results show that the simulation can provide very valuable information to detect interpret complex SAR images changes.
As an important complementarity of remote sensing technique,Unmanned Aerial Vehicle (UAV) is attractive to various applications such as target searching,island management,disaster monitoring and low-altitude photogrammetry.The quad-rotor UAV,Compared to fixed-wing UAV,has the advantages of Vertical-Taking-of-Landing (VTOL) and low-altitude flexible flight.However,the quad-rotor UAV has small-volume and light-weight and is more easily influenced by the wind,the acquired imagery maybe have bigger tile angle and more obvious geometry distortion will would result in more difficulty on the image feature matching and image mosaic. For this question,we apply the Scale Invariant Feature Transformation (SIFT) to image feature matching and Thin Plate Spline (TPS) transformation to automatic registration of the micro quad-rotor UAV imageries in this paper.The registration imagery based on TPS transformation is then compared to that based on the affine transformation and the polynomial transformation by evaluating the visual effect of the mosaic imagery and Root Mean Square (RMS) statistic,which shows that after precise SIFT feature matching,the registration RMS accuracy and the visual effect of mosaic imagery of TPS transformation are best,and satisfies the demand of rapid registration and mosaic of micro quad-rotor UVA imagery,since TPS transformation considered both rigid transformation and partial nonlinear distortion of the micro quad-rotor UAV imagery.
Because of its sharing and high time resolution,MODIS image has became the widely used data source in large area snow monitoring research.Fusion between MODIS image bands can provide a higher resolution image data source for snow study.In order to utilize the spatial and spectral information of MODIS 250m bands data effectively,this paper takes band 1 and/ or band 2 as the high resolution images and uses the remote sensing image fusion methods SFIM and wavelet transform to fuse bands 3~7.The fused image was input into the line spectral mixture model to extract snow area at sub-pixel scale.In addition,the accuracy of the extracted snow area was assessed by the Landsat TM snow maps,which was used as the “truth value”.The results show that the quantity accuracy of snow cover area gotten form fused image is 75% ,increased by 6% than that of generated form unfused image and band fuse of MODIS images is an effective way to extract high accuracy snow information.
SAR co-registration is one of the crucial steps in the process of SAR images analysis,especially for multi-temporal SAR images analysis.Focusing on the characteristics of the repeat-pass space-borne SAR image,an auto-registration method for repeat-pass SAR complex image pair is proposed based on a multi-steps,which combines the Fourier transform and correlation-coefficient algorithm in multi-scale scheme.Firstly,an initial offset is estimated in Fourier domain by using the 3rd level wavelet decomposition image.Then,taking these initial offsets into account,coarse and fine registration is applied to the whole image and the oversampled image successively by correlation analysis of the images on several windows distributed randomly.ALOS-PALSAR and Radarsat-2 data are used for validation of this registration method.And the experimental results show that the proposed technique achieves the co-registration of the SAR images with the accuracy up to a fraction of a pixel automatically without a priori knowledge and this approach is efficient.
HICO (Hyperspectral Imager for the Coastal Ocean) is the first hyperspectal imager on board International Space Station (ISS),and it is specially designed for coastal ocean remote sensing.In this paper,main characteristics of HICO are introduced and compared with another two on-orbit hyperspectral imager:EO-1 Hyperion,HJ-1A HyperSpectral Imager.In the case of turbid water environment of the Yellow River Delta,three typical kinds of objects are selected for comparasion of their radicane and ground reflectance after the atmospheric correction with FLAASH (Fast Line of Sight Atmospheric Analysis of Spectral Hypercubes).Results show that HICO’s spectrum curves are smooth and consistant with in-situ measured spectra,while HSI’s fluctuate strongly,particularly in blue bands.Water retrieval using ~820 nm band in atmospheric correction distinctly improves the reflectance accuracy around absorption band of water vapour.On the whole,HICO has better performance than HSI in delivering the spectral characteristics of objects in coastal zone.
In order to represent vague and imprecise value and improve the classification accuracy of remote sensing images,a fuzzy rule-based classification method was proposed.Firstly,by transforming the fuzzy membership function values into corresponding fuzzy subsets,fuzzy rule conditions were established.And then,the fuzzy rule database was derived from samples.Finally,based on the fuzzy nearness degrees of rule conditions were calculated from classified data and fuzzy rule database,the land use was classified.The experimental results show that the proposed method is able to significantly improve the classification accuracy than the maximum likelihood method while the data contains complex mixture of spatial information.Furthermore,this method can get the approximate results as the maximum likelihood method while the data contains relatively homogeneous spatial information.
Removing thin cloud/fog cover is one of regular tasks in remote sensing image processing.In this study,it is found that the effects from thin cloud cover to spectral signal usually have a performance of low-frequency characteristic and lead to both increased brightness V and reduced saturation S of the image signals.As the thickness of the cloud gradually changes from the center of cloud to the edge,the corrections for V and S also vary gradually.Therefore the relations between image distance D and S and between D and V divided by S were obtained by image sample and regression analysis.And then through a pixel-by-pixel correction in both S and V with the two relations,the spectral signals can be recovered well.When added a parameter of orientation angle,these relations can also be used to remove cloud in an asymmetric cloudy field.It has been demonstrated by several tests that the spectral signals can be recovered quite well with above mentioned method and has a low calculating costs for sampling,regression and correction.
In this paper,a real-time spectrum reconstruction system of Forman method based on GPU is proposed.Experiments show that the processing rate of the system is 2.82 times as high as the data rate of HJ-1A satellite and meet the demands of real time generation of primary products and full resolution display based on moving window.Forman method and Mertz method are compared based on GPU.In addition,the Forman method is analyzed in three different processing environments.These researches provide necessary data and methods to the construction of a practical system.
Taking the ETM+ data of Minqin oasis and the surrounding area as example,this paper analyses the roles of the spectral thematic index and texture characteristics in land cover classification,as well as the responses of different classifiers.It is not necessary that adding thematic index into the original band data can improve the overall classification accuracy.While adding texture variables at the same time the thematic index is able to fully play their roles.Maximum Likelihood Classifier(MLC) and Artificial Neural Network(ANN) response to the input variables differently.The former obtains the best results when three types of data all be involved and the latter obtains the highest overall classification accuracy when the original band be removed.The study shows that the spectral thematic index can improve land cover classification accuracy,but classifiers and variables (bands) combinations must be carefully chosen.
The surface feature of alluvial fans in arid areas is an important way to study the paleoclimate and paleoenvironment changes.The roughness of alluvial fans is an significant indicator of the evolution of alluvial fans.Synthetic Aperture Radar (SAR) has already been the key spatial observation tool to extract the surface features of alluvial fans automatically.However,the existing surface roughness retrieval models can hardly obtain good results.Using Radarsat-2 images,this paper carried out the research of roughness inversion of Shule River alluvial fan,and proposed a new roughness inversion method which is suitable for arid areas based on OH model.Through field measurement validation,the new method gained reliable results and solved the problem that original models can hardly gain good results in arid area.So this study provides a powerful tool to extract the surface features of the alluvial fans in arid areas.
Soil moisture is an important environmental factor and procedure parameter in fields of meteorology and climatology.Global and sequential products of satellite retrieval soil moisture were provided by AMSR-E.In this study,ECWMF and NECP reanalysis data and AMSR-E soil moisture were compared in time and space,the AMSR-E soil moisture was verified with consistency evaluation of three methods which mentioned above and the AMSR-E,ECWMF and NECP soil moisture in China were confirmed by observation data,such as soil moisture and precipitation.The result indicated that AMSR-E,ECWMF and the NECP soil moisture spatial distribution characteristics in globe or China have good consistency,but AMSR-E soil moisture was smaller than the others,especially when the value of the AMSR-E was low,the gap was bigger between them;Besides,the soil moisture and precipitation have perfect corresponding relation and there was better the corresponding relation between ECWMF and NECP relatively;Compared with the soil moisture of observation stations,ECWMF and NECP soil moisture was larger and AMSR-E soil moisture was smaller.Furthermore,the root mean square errors of ECWMF,NECP and observation station were all smaller than AMSR-E during 1 645 times data which acquired on a national scale.
The useful information provided by a single satellite is often not sufficient,for the use of multiresolution sensors from different satellites,we examined how well HJ-1A/B CCD sensor vegetation observations replicate Landsat TM/ETM+ vegetation observations.The study investigates the quantitative relationship between the vegetation indices of HJ-1A/B CCD and Landsat TM/ETM+ sensors,and more important,the difference in the vegetation observations between the two sensors.Three vegetation indices (the Normalized Difference Vegetation Index (NDVI),the Soil Adjusted Vegetation Index (SAVI) and the Enhanced Vegetation Index (EVI)) were calculated from both HJ-1A/B CCD and Landsat TM/ETM+ reflectance data.The quantitative relationship was revealed through regression analysis on the scatter plots of the vegetation index values for the same target of the two sensors based on three tandem images pairs.The models for the conversion between the two sensors were also obtained from the regression,then the accuracy of the obtained conversion models were evaluated based on the other two image pairs.The results show that there is a good agreement between the two sensor corresponding vegetation indices with the conversion models also have high accuracy,it will contribute to joint use of two sensors for vegetation monitoring by translated the vegetation index from one sensor to another based on the obtained conversion models.The relative spectral response function differences in the red and near infrared bands between the two sensors are believed to be the main factor which contvibute to their differences in vegetation observations.
Inversion of TN and TP’ concentration through the method of derivative and wavelength ratio on the basis of the hyper spectral data which was measured from April to October in 2005.The results indicate that differential method in the 1nm intervals 455 nm differential values of TN concentration in more sensitive,maximum Negative correlation coefficient is -0.857,maximum positive correlation coefficient is wavelength ratio’s(1 015/528 nm)0.803.Using these two factors as variables for multiple regression,the best inverse model of TN was got,the absolute error of model is 0.16,between measured and predicted values of R2 is 0.839.The band ratio (671/680 nm)of TP concentration is more sensitive,the maximum Negative correlation coefficient is -0.889,because the variables are selected as a second polynomial regression,the best inverse model of TP was got.The absolute error of the model is 0.003,between measured and predicted values of R2 is 0.934.
An improved Carnegie Ames Stanford Approach model due to remote sensing data and climate variables were used to estimate the Net Primary Productivity (NPP) of Xuzhou in the June of 2006,2008,2010.The results of analysis variations and the relationship with climatic factors were as the follow:The mean NPP of terrestrial vegetation in Xuzhou showed decreasing trend in recent years because of the changes in climate and environment;The NPP values were higher in most of places where woodland,lawn and farmland were aggregation,the variations of NPP might be significant in different vegetation types,the NPP of farmland was more than lawn and woodland in 2006 and 2008,and the NPP of farmland was more than farmland and lawn in 2010.Based on the analysis of the correlation and partial correlation between NPP and climatic factors (temperature,precipitation and sunshine duration),it could be concluded that the influences of climate were complicated with NPP,and the dominant factor was not constant in this area.Because water and heat was in a good condition for vegetation growth in 2006 and 2008,the relationship between sunshine duration and NPP was more stronger;The temperature and precipitation became the limiting factors in 2010 because of lower in themselves and longer in sunshine duration.The different requirements of the different vegetation types in light,heat and water could be reflected by the variations of correlation and partial correlation between NPP and climatic factors in different vegetation types.
Leaf Area Index (LAI) and Extinction Coefficient (EC) of vegetation canopy are parameters that represent vegetation canopy structure and the earth underlying surface of light energy utilization.There are lots of research and application about the retrieving of LAI at home and abroad,however,the study about the retrieving of EC is rare.The southern of china is located in the humid subtropical region with abundant rainfall,mixed vegetation and different density.In order to better match the measured data of LAI and Photosynthetic Available Radiation (PAR),and better reflect the condition of mixed vegetation and different density,This paper used the Landsat ETM image as the remote sensing data,and compared the image-fusion methods of HSV,Brovey and Gram-Schmidt(GS)to choose the most suitable image-fusion methods to fuse the ETM image into multi-spectral data with 15 m spatial resolution.This paper selected the Meijiang watershed which is the headstream of Poyang Lake as the study area,based on the measured data of LAI and PAR of dominant vegetation,used the empirical formula method of vegetation index to retrieve the LAI and based on the Beer-Lambert law,and establish the retrieving models of EC of dominant vegetation canopy in the watershed using four vegetation indices,namely NDVI,SAVI,RVI and TSAVI.The results show that the NDVI was the most robust index with an R2 value of 0.81 for the estimating of LAI,but with an R2 value of 0.68 for the estimating of EC.Moreover,the spatial distribution of the LAI and EC was retrieved too,which can provide important input data for the SWAT model.
Typical land surface spectrum data are mainly applied in related basic research of quantitative remote sensing and image classification.We selected representatively spectral library set of vegetation spectrum(73 items),manmade spectrum(100 items)and soil spectrum(30 items) from USGS and JHU spectral library,and analysed typical separability features of materials spectral features in 4.2~2.5 μm wavelength range within a hierarchical classification scheme,Application of Bhattacharyya distance to quantitatively calculated among different categories objects spectrum spectral differences:the calculation results show that spectrum separability metal,brick and concrete manmade material spectrum for vegetation,soil and other natural materials spectrum have greater spectral differences,and separability of plastic between natural features is smaller.Additionally,an evaluation of the most suitable wavelengths for separation of spectral library set identified specific spectral features that provided the best separation.Based on the statistical characteristics of spectral differences could reflect the optimal band.The study provides a basic knowledge reference of spectral discrimination analysis in a variety of material spectrum and also have the certain reference significance of remote sensing image land classification in a larger scale.
The canopy hyperspectral reflectance data of four typical kinds of wetland vegetation (Phragmites australis,Cyperus malaccensis,Kandelia candel and Spartina alterniflora) was acquired by Analytical Spectral Devices at the Shanyutan wetland of Minjiang river estuary in October 2011.The data was analyzed by adopting Viewspecpro and Oragin8.0 software to research the effects of band position and bandwidth on NDVI values.The following results showed that band positions and bandwidths of red band and near-infrared (NIR) band had significant different influences on NDVI values of four kinds of wetland vegetation relatively.
(1)When the red wavelengths fixed,NDVI of all vegetation were increased rapidly along with the increase of band center position,while NDVI appeared a small valley value in 925~1 050 nm.P.australis and S.alterniflora showed this valley value more obviously.Moreover,the different bandwidths influences on NDVI were that,in 725~825 nm,the lower NDVI values came out with the wider bandwidths; in 825~925 nm,the bandwidths had less influence on NDVI; in 925~1 100 nm,the narrower bandwidths got greater influences on NDVI.(2)When the NIR wavelengths fixed,the NDVI values of K.candel and S.alterniflora were increasing gradually with the increase of band center position in 600~680 nm,but P.australis and C.malaccensis did not showed this trend.In 680~700 nm,the NDVI of four kinds of wetland vegetation decreased rapidly.Bandwidths influences on NDVI were different with red wavelengths fixed condition.In 650~700 nm,the lager NDVI values came out with the narrower bandwidth; the bandwidths had less influences on NDVI in 600~650 nm.(3)The bandwidth of red band influences on NDVI values of four kinds of wetland vegetation is greater than the bandwidths of NIR band.
With the rapid development of remote sensing technology,it is possible to observe sea surface temperature (SST).As one of the major methods of observing SST,microwave remote sensing by satellite has received increasing attention.The course of development on microwave remote sensing for SST observations and the advantages compared with infrared observations are summarized,basic retrieval theory and two major retrieval algorithms:Multiple Linear Regression Algorithm and Non-Linear Iterative Algorithm are concluded,moreover,operating parameters and the channel characteristics and other advanced microwave radiometers carried on of TMI,AMSR-E FY-3 and HY-2 are introduced with the retrieval algorithms for each sensor,finally,the effective factors of SST retrieval are analyzed,the problems at present are listed,and subsequent studies are prospected.
Due to the limitation of spatial resolution,there are lots of mixed pixels in spaceborne hyperspectral images.Spectral unmixing of hyperspectral images is an important premise for accurate terrain classification and identification.Compared with traditional spectral unmixing techniques based on the linear mixing model,nonlinear spectral unmixing techniques has better performance in finding endmembers and their abundances.The principle of nonlinear spectral mixture is analysed,and nonlinear unmixing algorithms increased in recent years are summarized.This paper emphatically introduces bilinear model,neural networks,nonlinear spectral decomposing based on kernel function and manifold learning.Some future directions of research are introduced.
This paper concluded two different kinds of conception of the albedo,and two main method of measurement,analysed the element that effect the albedo,such as snow grain size,water content,soot,snow density,snow depth,solar zenith angle,atmosphere condition and new snow,then described the universal method to retrieve the albedo from remote sensing,including radiometric calibration,atmospheric correction,anisotropic correction and narrowband to broadband albedo conversion.And we finally summarized the research and application of albedo.
