YAN Mei-chun. Research and Contrast on Several Vegetation-Classification Methods of High-Resolution Satellite Image Data[J]. Journal of Remote Sensing, 2007, (2): 235-240. DOI: 10.11834/jrs.20070232.
Research and Contrast on Several Vegetation-Classification Methods of High-Resolution Satellite Image Data
Texture information can avoid the matter of ’same spectral from different materials’ and ’same material with different spectral’ which must be faced with when making classification with only spectral information of high resolution image
but few people use only texture information to distinguish different vegetation of high resolution image.This paper takes the Zhongshan scenic region of the eastern side of Nanjing City as the study area
and uses the satellite data of IKONOS to explore the methodology of discriminating the different types of vegetation.When texture information is employed
four texture statistics such as: CON(contrast)
COR(correlation)
HOM(homogeneity)
and MCON(make-up contrast) are used.In order to test the result
Vegetation index information is used as well as spectral information
three VI such as: NDVI
MSAVI and SAVI(whose values of L equal to 0.5 and 5 respectively)are used to decide the best one who stands for some vegetation type
and SAVI5(value of L is 5)does well.A threshold is selected in texture and VI information to discriminate every type.From spectral information two methods are used: minimum distance supervised classification and ISODATA unsupervised classification.When processing the image
some steps must go: firstly
recovering the IKONOS image data;secondly
discriminating the vegetation types as: grassland
garden
conifers
mixed forest and bamboo
wide-leaf forest using the different methods
including the discrimination based on the pixel spectral data
the discrimination based on vegetation index
and the discrimination based on minor texture on the images.Through comparing all results the following conclusions have been drawn: texture information has the best classification accuracy
and the VI has better result
and the spectral information only has the worst
from which the result of unsupervised classification is the lowest.We can conclude that texture information can discriminate vegetation types and other material perfectly because it response to the spectral composition and properties of some material.
Shaoqiang WANG 中国地质大学(武汉),地理与信息工程学院,区域生态过程与环境演变实验室;中国科学院地理科学与资源研究所,生态系统网络观测与模拟重点实验室;中国科学院大学,资源与环境学院
相关机构
China University of Geosciences (Wuhan), School of Geography and Information Engineering, Laboratory of Regional Ecological Processes and Environmental Evolution
Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, State Key Laboratory of Resources and Environmental Information System
Research Center for UAV Applications and regulation, Chinese Academy of Sciences
Key Laboratory of Low Altitude Geographic Information and Air Route, Civil Aviation Administration of China
Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Key Laboratory of Ecosystem Network Observation and Modeling