IJBTT

Analysis of Dimensionality Reduction Techniques for Hyperspectral Image Classification

	International Journal of Engineering Trends and Technology (IJETT)
	© 2015 by IJETT Journal
	Volume-21 Number-2
	Year of Publication : 2015
	Authors : Sinduja.R, Prof.S.Chidambaram, Dr.A.Sumathi
	DOI :  10.14445/22315381/IJETT-V21P219

Citation 

Sinduja.R, Prof.S.Chidambaram, Dr.A.Sumathi"Analysis of Dimensionality Reduction Techniques for Hyperspectral Image Classification", International Journal of Engineering Trends and Technology (IJETT), V21(2),111-115 March 2015. ISSN:2231-5381. www.ijettjournal.org. published by seventh sense research group

Abstract

Hyperspectral imagery utilized in remote sensing applications provides richer data concerning materials than multispectral imagery. The new larger information volumes from hyperspectral sensors gift a challenge for ancient process techniques. as an example, the identification of every ground surface picture element by its corresponding spectral signature remains tough as a result of the large volume of knowledge. standard classification strategies might not be used while not dimension reduction pre-processing. this is often as a result of the curse of spatiality, that refers to the very fact that the sample size required to estimate a perform of many variables to a given degree of accuracy grows exponentially with the quantity of variables. Principal part analysis (PCA) has been the technique of alternative for dimension reduction. However, PCA is computationally high-ticket and doesn't eliminate anomalies which will be seen at one arbitrary band. The high-dimensional nature of the information collected by such sensors not solely will increase procedure complexness however can also degrade classification accuracy. to deal with this issue, spatiality reduction (DR) has become a crucial aid to rising classifier potency on these pictures. Dimension reduction algorithms don't scale back the dimension of knowledge with the goal of reconstructing AN approximation to the initial signal. Instead, they ask for a minimal illustration of the signal that sufficiently retains the requisite data for prosperous unmixing within the lower dimension. Dimension reduction algorithms are designed to attenuate errors within the procedures performed within the lower dimension.

References

[1] An Investigation On Hyperspectral Image Classifiers For Remote Sensing, Okan Bilge Özdemir, June 2013
[2] An Unsupervised Spectral Matching Classifier Based On Artificial Dna Computing For Hyperspectral Remote Sensing Imagery Hongzan Jiao, Yanfei Zhong, Member, Ieee, And Liangpei Zhang, Senior Member, Ieee2013
[3] Automatic Reduction Of Hyperspectral Imagery Using Wavelet Spectral Analysis Sinthop Kaewpijit, Member, Ieee, Jacqueline Le Moigne, Senior Member, Ieee, And Tarek El-Ghazawi, Senior Member, Ieee, Ieee Transactions On Geoscience And Remote Sensing, Vol. 41, No. 4, April 2003
[4] Keinosuke F., Raymond R.H., (1989), “Effects of sample size in classifier design”, IEEE Transactions on Pattern Analysis and Machine Intelligence, 11873885
[5] G. Hughes, (1968), “On the mean accuracy of statistical pattern recognizers”, IEEE Transactions on Information Theory
[6] Hall M. A. and Smith L. A. (1997). “Feature subset selection: A correlation-based filter approach”. Proc. Int. Conf. Neural Inf. Process. Intelligence Information System (pp. 855-858).
[7] Peng H., Long F., and Ding C. (2005). “Feature selection based on mutual information: Criteria of max-dependency, max-relevance, and minredundancy.” IEEE Trans. Pattern Anal. Mach. Intel, (pp. 1226-1238).
[8] Anil K. J., Robert P.W. D., Jianchang M. (2000). “Statistical Pattern Recognition: A Review.”, IEEE Transactions on Pattern Analysis and Machine Intelligence
[9] B. SchÄolkopf, A. J. Smola, K. R. Muller. (1998). “Nonlinear Component Analysis as a Kernel Eigenvalue Problem”, Neural Computation.
[10] Craig R., Jie S. (2002). “Principal Component Analysis for Hyperspectral Image Classification”, Surveying and Land Information Science, (pp. 115-122).

Keywords
PCA is computationally high-ticket and doesn't eliminate anomalies which will be seen at one arbitrary band.