Best Band Selection using Principle Component Analysis Algorithm from Remote Sensing Data

The best band selection from remote sensing image plays an important roles in multispectral and hyperspectral remote sensing image processing due to the intercorrelation that inherent in the multispectral images taken by remote sensing sensors.
In this paper we use principle component analysis algorithm applied on remote sensing data and find covariance matrix for bands that should be processed then find eigen vector using Jacobi methods .The algorithm was applied on multispectral images of Thematic Mapper sensor , it concluded that the six band was the best band , the value of its eigen value was the biggest one and the value of signal to noise ratio equals to 74.7217. This algorithm is constructed using Visual C# 2008 that is characterized by efficient and high speed implementation.