期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2014
卷号：3
期号：7
页码：2394-2400
出版社：Shri Pannalal Research Institute of Technolgy
摘要：It is common to represent the remote sensing images in true and false color composites in various parts of electromagnetic spectrum. Visualising color images provide a crude view of the terrain and that too at the three selected wavebands. In order to derive detail and accurate information about the earth surface, more spectral information contained with an extended range of wavelengths is necessary. Hyperspectral image cube contains of different ground truth objects which have numerous, narrow and in close proximity spectral band information centering on equally distributed wavelengths ranging from visible to near infrared spectrum. This narrow wavelength helps in identifying various vegetation classes, mineral types, small military targets and many more objects on the surface of earth. This paper presents existing conventional methods, issues, and their limitations of information extraction from hyperspectral imagery dataset. This review helps that designing and implementation of a suggestible digital image analysis approach is a prerequisite to get a better classification accuracy of optically sensory data into a thematic map layer i.e., from dimensionality reduction, perpixel, subpixel to super resolution mapping (SPM) based on the spatial dependences of fractional abundances and the anomaly detection is the case when one ca not know the signature member of the target which try to find pixels that deviate from the background. exact use of various feature information of optically sensory data and the usage of best required approach of classification method are effectively significant for producing better classification accuracy.
关键词：Perpixel; Subpixel Classification; Super Resolution ; Mapping; Hyperspectral Dataset; Colorimetry and Advance Image ; Processing Techniques
