*** GUI-based HighView software for personal use free of charge since 10/2011 ***

Image fusion is a concept of combining multiple images into composite products, through which more information than that of individual input images can be revealed.

As an example of image fusion, pan-sharpening describes a process of transforming a set of coarse (low) spatial resolution multispectral (colour) images to fine (high) spatial resolution colour images, by fusing a co-georegistered fine spatial resolution panchromatic (black/white) image. Typically, three low-resolution visible bands - blue, green and red - are used as main inputs in the process to produce a high-resolution natural (true) colour image. Figure 1 illustrates this: the middle image is a natural colour image with a spatial resolution of 2.4 m (resampled 400%), and the left image a panchromatic image with a spatial resolution of 0.6 m; by combining these inputs, a high-resolution colour image is produced. In the fused output, spectral signatures of the input colour image and spatial features of the input pan image, as the best attributes of both inputs, are (almost) retained. The spectrally and spatially enhanced image is often visually appealing, starting to compete with high-resolution aerial photographs.

Figure 1: Image pan-sharpening with QuickBird images. The left and middle input images are obtained from DigitalGlobe, and the right-hand image is the fused result from HighView. All images are subject to the same histogram stretch.

Figure 2: Image pan-sharpening with IKONOS images. The left and middle input images are obtained from Space Imaging, and the right-hand image is the fused result from HighView.

The past few decades have seen quite a few image fusion and pan-sharpening methods in the public domain, including those based on multi-resolution wavelet transforms, PCA (Principal Component Analysis) transforms, and IHS (Intensity-Hue-Saturation) transforms. These methods, however, largely disregard important spectral characteristics of specific satellite sensors, therefore no consistent, colour-preserving results can be achieved. In other words, a simple data-driven approach without referring to spectral evidence can hardly produce satisfactory outcomes. This becomes more apparent when the recent generation of high- and medium-resolution satellite images is available, where there are marked spectral disparities between colour bands and the panchromatic band.

For more information on spectral features and curves of recent satellite sensors, please refer to websites of major image providers:

http://www.astrium-geo.com/pleiades/ (Pléiades)

www.digitalglobe.com (QuickBird, WorldView-2)

www.geoeye.com (IKONOS, GeoEye-1)

Korea Aerospace Research Institute (KOMPSAT-2)

Taiwan National Space Organization (FORMOSAT-2)

Japanese Advanced Land Observing Satellite (ALOS, or "Daichi")

www.spot.com (SPOT-5)

landsat7.usgs.gov (Landsat ETM+)

http://landsat.usgs.gov/ (Landsat 8)

HighView searches physical evidence of spectral characteristics and applies optimisation methods to perform ideal image pan-sharpening. While the fused imagery at a finer spatial resolution still retains the mean of the coarse-resolution input, its standard deviation should naturally become larger due to  increased details and heterogeneity of image features at finer resolutions (Figure 3). These important characteristics set HighView apart from other pan-sharpening methods on the market that incorrectly claim the preservation of standard deviation of the fused imagery at a finer resolution.   

Figure 3: Statistical changes from coarser-resolution input to fused, finer-resolution imagery in HighView.

 

ACKNOWLEDGEMENTS:

All Pléiades, WorldView-1, QuickBird, GeoEye-2, IKONOS, SPOT-5 and Landsat ETM+ images used at GeoSage website are demonstration, sample, or free images from Astrium, DigitalGlobe, Space Imaging, SPOT Image, NASA, USGS, GeoGratis and GLCF. The following banner picture is created by http://wordle.net/