Dissertation
Dissertation > Astronomy,Earth Sciences > Surveying and Mapping > Photogrammetry and Surveying, Mapping and Remote Sensing > Surveying, Mapping and Remote Sensing technology

Knowledge-based Remotely Sensed Imagery Classification Method Research

Author WangHuiLin
Tutor LiuYong
School Lanzhou University
Course Cartography and Geographic Information Systems
Keywords Arid and semi-arid areas Remote sensing image classification Tengger Desert in southern Accuracy Assessment
CLC P237
Type Master's thesis
Year 2007
Downloads 632
Quotes 2
Download Dissertation

To improve arid , semi- arid regions of land use / cover automatic classification accuracy as the goal , around to study remote sensing image classification method based on knowledge of the process . First, determine the land use / cover classification scheme ; then starting from the visual interpretation of the principle of the visual characteristics of the study area, land use / land cover types in remote sensing image and spectral characteristics , texture characteristics , spatial distribution characteristics of the study area the when phase characteristics of the analysis; then from the expression of knowledge , knowledge base construction is discussed based on remote sensing image classification method based on knowledge : the last of the study area, a knowledge - based remote sensing image classification , the classification results evaluation. Above , we get the following conclusions : (1) knowledge - based remote sensing image classification method is to increase the surface complexity of drought , one of the effective ways to semi-arid areas of land use / land cover classification accuracy automatically . Although the study area, surface land use / cover complex and the lack of the same period , the actual validate the data , there is a certain deviation of classification accuracy evaluation . But overall , due to a combination of the spectral characteristics of the remote sensing image texture features , spatial distribution characteristics and phase characteristics , based on knowledge of remote sensing image classification accuracy than traditional supervised classification methods have improved to some extent . (2 ) for the arid and semi-arid regions , texture features , spatial distribution characteristics and phase characteristics can better improve the classification accuracy . The texture features highlights morphological differences desert and Gobi information . Spatial distribution characteristics of major enhancements mountain shrub and mountain forests , mountain meadows and farmland has obvious spatial distribution of the surface feature differences , to reduce their mixed points . When the phase characteristics distinguish farmland growing crops and other desert . (3) Based on the knowledge of the classification method can be effectively consolidated various other auxiliary data for improving the accuracy of the classification , the auxiliary data itself is the quality and the auxiliary data between the remote sensing image registration accuracy, to some extent will affect the classification precision. Therefore, in the practical application must pay attention to a precise alignment between the auxiliary data selection and different data .

Related Dissertations
More Dissertations