Data integration using color space transforms

The demand for increased spatial resolution without sacrificing spectral discrimination can be fulfilled by integration of data from different sensor systems and satellite programs. Data of high spatial resolution are frequently available in panchromatic (black-and-white) form rather than multispectral. Techniques gave been developed to combine the higher resolution panchromatic data with a multispectral data set of lower spatial resolution.

The standard method of integration modulates the intensity of the mutispectral with the panchromatic data. A less subjective approach uses an algorithm that describes color in terms of intensity (I), hue (H), and saturation (S).

Combinations of high resolution panchromatic data (SPOT panchromatic) and lower resolution multispectral data [Landsat thematic mapper (TM), SPOT XS] have been developed. The SPOT data were acquired on April 3, 1986, and the Landsat TM data were acquired on April 5, 1986. The data sets were registered to each other and the multi-spectral data sets were contrast enhanced. The enhanced multispectral data sets were then transformed from red/green/blue (RGB) color space into IHS space. In each case (TM/SPOT panchromatic and SPOT XS/SPOT panchromatic), the SPOT panchromatic data were remapped on a cumulative histogram percentage basis to match the multispectral "I" data. These remapped SPOT panchromatic data were substituted for the original multispectral "I" and the hybrid IHS data transformed back into RGB space for display.

While this technique is experimental and still being refined, the results, to date, indicate that the IHS method will be valuable for generating improved images that effectively present both high resolution spatial digital data and multispectral data.