Research On Remote Sensing Thematic Information Extraction In Alpine Region And Development Of Query System

With the rapid development of remote sensing technology in recent years,the remote sensing interpretation has been gradually applied to various fields.This study is based on the remote sensing integrated survey project whose purpose is to extract the thematic information on the geology,engineering,hydrology,geomorphology and cover of the alpine region of Qinghai Tibet Plateau.The extraction method for different thematic information in Alpine Region was researched by analysing the image data.And then the high accuracy thematic map was drawn though the original geological,mineral,terrain,geological disaster information and high-resolution remote sensing image data.On this basis,the knowledge map inquiry system for the research area that allows users to management and use of thematic interpretation of the results easier was further developed.The process of remote sensing Thematic Interpretation mainly depend on series of the latest high-resolution satellite images,United States Landsat-8 satellite images,DEM data and other auxiliary data.Though the image preprocessing of geometric correction,atmospheric correction,image fusion and image enhancement,it was realized that the correction of the geographical coordinates,the unity of the pixel value information and the combination of the image spatial resolution and spectral resolution of remote sensing image.At the same time,the method of image enhancement was researched in order to extract the target information,which is helpful for the further interpretation.Because of the special characteristics of the landform,hydrology and cover in the alpine region of the Qinghai Tibet Plateau,most of the study area is exposed rock mass,Gobi,sandy land that in a single color.Therefore,the traditional supervised,unsupervised classification and other simple methods based on spectral information is difficult to achieve accurate identification and classification of features.This study chose object-oriented classification as the main method to interpret the translation work.The information extraction method of object-oriented segment the remote sensing image into different sizes objects through different methods first,then extract these objects feature information.This method make full use of the spectral features and spatial characteristics of size,shape,texture of the remote sensing images to effectively compensate for the disadvantages of classification based on pixel level,and improving sort precision to a certain extent.This study used two ways which based on rules and based on the sample selection for target extraction,and tested the accuracy of interpretation that reaches 0.87 on the overall Kappa value.The methods to eliminate the interference of the mountain shadows in process of lake water extraction in the high mountain areas was further researched.Although the spectrum-based methods such as spectral correlation or spectral threshold controlling may partially remove the shadows.But the real conditions of water body mapping were rather complex,and the above-mentioned methods may underestimate or overestimate the lake water extents based on different thresholds,even missing some lakes.The method proposed in this study firstly uses the spectral feature of water to extract the potential complete lakes,then remove the shadows by differentiating the relationships of spatial position of lakes water,mountain shadows and valleys.As a result,we cannot only keep relative complete lakes,but also reduce the interference of mountain shadows efficiently.By adjusting the parameter for minimal size of lake extraction,we can remove more than 90%of the mountain shadows on the premise of saving relatively complete lakes in the high mountain areas.The final results of thematic map is difficult to the manage and use because of the large range of project area(about 150 000 Km2),the complex interpretation content which including geology,hydrology,geomorphology,cover etc,and thematic map that contains many categories and attributes.Therefore,this study developed the knowledge map query system with the thematic map of interpretation results to make the management and operations of the project result easier.The development of query system based on the Visual Studio.NET framework and ArcEngine development platform,realized the functions of thematic map view,selection,attribute query with C#language.Finally,the user can manage the interpretation information better to understand the study area systematically and obtain useful information.