Correspondence analysis on Landsat TM remote sensing image

Taking Mentougou district in the west of Beijing for example, the authors discussed Correspondence Analysis (CA) on TM remote sensing image. The experimental results showed that, CA developed on the basis of Principal Component Analysis (PCA) involved more than 98% information from the original TM image. Thereinto, the first component covered upwards of 88% information. The first component almost concentrated entirely useful information, while the rest components were mainly noise information. The intensity component of CA better represented the brightness values throughout the whole data compared to PCA. In addition, CA not only considered the relationship among bands, also the correlation between each band and surface features, which definitely expressed the physical meaning of each component and made up the limitation of PCA.