Font Size: a A A

Wetland Information Extraction Based On Remote Sensing Imagery And Statistical Analysis Methods

Posted on:2015-01-18Degree:MasterType:Thesis
Country:ChinaCandidate:T MengFull Text:PDF
GTID:2250330428490787Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Wetland landscape is the natural complex by the interaction of water and land form, isan important part of the earth’s ecological environment, and also is one of the importantcontents of geography census together with forest, grassland, desert, and ocean. With therapid development of remote sensing technology and remote sensing technology, satelliteremote sensing imagery has become the important resource for fast and accuratemonitoring the land use change. The wetland landscape is almost locating the border ofland and water and the environment of aquatic, semi aquatic, the related remote sensingimages should be influenced by "the same object with different spectrum, and the differentobject with the same spectrum". Therefore, wetland monitoring information extraction andstatistical analysis based on high resolution remote sensing image has become one of thecurrent hot research topics in remote sensing.Taking Zhalong Nature Reserve as the experimentation area, and high resolutionremote sensing data as the data source, this paper studied the effect of differentclassification algorithms to extract wetland monitoring information, and evaluated andcompared the classification effect by multiple indexes. It provided important data resourcesfor the planning, strategic decisions and land use analysis for Zhalong Nature Reserve. Themain research contents of this paper are as follows:(1) On the aspect of image classification and information changes detecting method,this paper studied the method and the process of object-oriented image classificationmethod in farmland, wetland, and other surface coverage information in order toeffectively access to land cover information data. By using the SVM method and theobject-oriented method, this paper implemented the farmland and wetland informationextraction for the experimentation area, and realized the change monitoring by comparingtwo-period data.(2) On the aspect of statistical analysis of geography census information, using thesoftware of ArcGIS, Excel, SPSS, system dynamics software, based on the land cover data,combined with social economy information and thematic information data, this paper established the mathematical model, and implemented the basic statistics, the transferringcharacter analysis of land cover information, and driving force analysis of land coverinformation change of geographical condition monitoring data for the experimental area,and analyzed the information change trend.
Keywords/Search Tags:remote sensing imagery, data preprocessing, wetland, object-oriented classification, SVM, information extraction
PDF Full Text Request
Related items