2. Fusion of hyperspectral data using segmented principal components transformation for enhanced colour display is proposed by means of partitioning the hyperspectral bands prior to Principal Components Transformation. Three methods are proposed, each one employing a different partitioning criterion. One of them uses a matched filtering procedure based on the spectral characteristics of various materials and is very promising for classification. The information content of the hyperspectral bands, as well as the quality of the obtained RGB images are assessed using existing measures. The obtained RGB representations are perceptually very good since they incorporate all characteristics of natural colour images and furthermore are very promising for classification purposes.

 

 

  • The partitioning of the bands depends on the spectral signature of certain materials.

  • The spectral signature of each material is used as a matched-filter H(ë), to emphasize hyperspectral bands at which this specific spectral signature has the highest value.

 

 

 

 

 

 
Spectral signatures of alunite (red line),
buddingtonite (green line) and kaolinite (blue line)