HighView is based on GeoSage's state-of-the-art and highly efficient image fusion algorithms that have been developed over the past few years and has been extensively tested with a variety of imagery sources.

(1) DOS-based HighView (Version 3.0, the latest version of HighView) includes the following utility programs:

- Advanced Image Fusion and Pan-Sharpening Tool (Fuse.exe)

- RGB Image Stretching Tool (RGB_Stretch.exe)

- RGB Image Composition Tool (RGB_Make.exe)

Product Brochure (PDF, size ~300KB)


Usage: Fuse.exe

Fuse.exe
<optimization_method> <output_image_type> -e
<enhanced_spatial_filter_option> <enhanced_green_color_option> <enhanced_pixel_offset_x_option> <enhanced_pixel_offset_y_option> -s
<image_stretch_band1_left_cut> <image_stretch_band1_right_cut> <image_stretch_band2_left_cut> <image_stretch_band2_right_cut> <image_stretch_band3_left_cut> <image_stretch_band3_right_cut> <nonlinear_stretch_band1> <nonlinear_stretch_band2> <nonlinear_stretch_band3>
<in_PAN.tif> <in_MS.tif> <out_PanSharpened.tif>


Parameters - Specific to algorithms

<optimisation_method>
0 - local optimization; 1 - global optimization

<output_image_type>
0 - true-color composite, e.g., QuickBird MS Bands 3/2/1;
1 - color-infrared composite, e.g., QuickBird MS Bands 4/3/2.

<enhanced_spatial_filter_option>
0 - none; 1 - detail; 2 - sharper

<enhanced_green_option>
Usually this value should be set to zero. If the user wishes to highlight vegetation in TOTAL green based on NDVI thresholds, set this value within the range (0.0-1.0) with a typical value of 0.2; i.e., vegetated areas with NDVI greater than 0.2 will be shown in TOTAL green. If the user just wishes to highlight vegetation GREENER, set this value within the range [1-10], typical value 4.0. If the output image type is set to 1 (color-infrared), this value should always be zero.

<enhanced_pixel_offset_x_option>
Usually this value should be set to zero. This option is to slighly adjust horizontal pixel shifts between PAN and MS inputs, when necessary. Rare usage recommended. For QuickBird and IKONOS inputs directly from data vendors, set this value to zero. If slight pixel mismatches exist between PAN and MS inputs, the user may set this value within the range [-0.25, 0.25]

<enhanced_pixel_offset_y_option>
Usually this value should be set to zero. This option is to slighly adjust vertical pixel shifts between PAN and MS inputs, when necessary. Rare usage recommended. For QuickBird and IKONOS inputs directly from data vendors, set this value to zero. If slight pixel mismatches exist between PAN and MS inputs, the user may set this value within the range [-0.25, 0.25]

Examples/Tips

1. This tool targets QuickBird and IKONOS imagery DIRECTLY supplied by
DigitalGlobe and GeoEye data vendors. For the majority of cases, both
panchromatic and multispectral inputs are 16-bit GeoTIFF files, and
meet the EXACT image size ratio of 4:1. Maximum image size for PAN
band is about 36500 x 36500 pixels, and max file size for output is 4GB.

2. Before pan-sharpening, users may wish to perform simple band combination and produce R/G/B true-color composite using the multispectral input:

RGB_Stretch.exe 3 2 1 -s 2 1 2 1 2 1 0 0 0 in_MS.tif in_MS_RGB.tif

3. For QuickBird imagery, typical pan-sharpening commandlines are as follows:

Fuse.exe 0 0 -e 2 0 0 0 -s 2 1 2 1 2 1 0 0 0 in_PAN.tif in_MS.tif out.tif

Fuse.exe 1 0 -e 2 0 0 0 -s 2 1 2 1 2 1 2 2 2 in_PAN.tif in_MS.tif out.tif

4. For IKONOS imagery, since multispectral data are usually provided with four separate bands - blue, green, red and near-infrared, users need to combine all multispectral bands together in order to make a new, four-band multispectral input file first, using RGB Image Composition Tool included in the package, i.e.,

RGB_Make.exe blue.tif green.tif red.tif nir.tif in_MS.tif

Then, use this tool for image pan-sharpening, e.g.,

Fuse.exe 0 0 -e 1 0 0 0 -s 2 1 2 1 2 1 2 2 2 in_PAN.tif in_MS.tif out.tif

5. To highlight vegetation in TOTAL green based on NDVI threshold, use
Fuse.exe 0 0 -e 0 0.2 0 0 -s 2 1 2 1 2 1 0 0 0 in_PAN.tif in_MS.tif out.tif

Just to highlight vegetation greener, use

Fuse.exe 0 0 -e 0 4 0 0 -s 2 1 2 1 2 1 0 0 0 in_PAN.tif in_MS.tif out.tif

6. To produce a color-infrared output, use

Fuse.exe 0 1 -e 2 0 0 0 -s 2 1 2 1 2 1 0 0 0 in_PAN.tif in_MS.tif out.tif

Figure 1: Demonstration of program options (DOS-based HighView Ver. 3.0)

 

Figure 2: Spatial enhancements in HighView

 

(2) GUI-based HighView (Version 2.5, an old version) is still available and ideal for casual users. It has two versions: General version and Professional version. Details are as follows:

User Guide (PDF, size ~514 KB).


Image band combination and stretching

• Linear histogram stretching: Users can set any % of total image pixels to be excluded in the stretch range.
• Nonlinear stretching: Users can adjust the level to display the full range of image intensities from very dark to white, and to make colour composites photographic for better visualisation by human eyes.
• Output imagery in 24-bit BMP and/or GeoTIFF format.
• A GUI-based batch processing mode allowing users to process hundreds or thousands of images overnight.
  
  
• Advanced simulation methods for making high-quality, natural-colour composite from false-colour bands: applicable to the Green/Red/NIR bands of SPOT-5, Landsat ETM+, ASTER, or any other generic colour infrared imagery (CIR). (Six colour templates are only included in HighView Professional Version 2.5. Or refer to a standalone utility program Spectral Transformer Fc2Tc-CIR)
  
  

Many image analysis programs apply linear and nonlinear stretching methods separately; however, the combined use of these two methods for individual R/G/B bands often results in output images with stronger visual appeal.

GUI 1: Band combination

GUI 2: Two image stretching methods

Pan-sharpening modules

• Local spectral/spatial optimisation method
• Global spectral/spatial optimisation method
• An efficient, improved bilinear resampling scheme
• Spatial enhancements (sharpen, detail, or none) for any selected set of bands
• Colour enhancements by adjusting R/G/B factors
• Option for matching histogram stretches used during band combination
• For Landsat ETM+ inputs, four options of spatial enhancements (including two for noises reduction)
• Pan-sharpened output imagery with full, unstretched dynamic range
• Pan-sharpened output imagery in 24-bit BMP format under linear and nonlinear stretches and/or GeoTIFF format
• An efficient batch processing mode that can be easily integrated into enterprise workflows.
• Applicable to QuickBird, IKONOS, Landsat ETM+ imagery and many others
• Applicable to increasingly popular SPOT-5 imagery (image fusion at spatial resolutions of 5m and 2.5m, including the simulation of true-colour composite from false-colour input bands) (Only included in HighView Professional Ver. 2.5)
• Option for either intensive or low memory usage in processing large size inputs (maximum image size of 32,768 by 32,768 pixels)
• If the file size of fused multi-band output is close to or larger than 2GB, HighView automatically produces outputs with separate bands

GUI: Main pan-sharpening module for 16-bit inputs

GUI: Spatial and colour enhancements for 16-bit inputs

GUI: Spatial enhancements for Landsat ETM+ inputs

Log file

 

(3) Comparison

	DOS-based HighView	GUI-based HighView
Version	3.0	2.5.3
Input	Dedicated to rapid processing of 16-bit QuickBird and IKONOS GeoTIFF files directly from DigitalGlobe and GeoEye	Broad with 8- or 16-bit data in ENVI, IDRISI, or GeoTIFF format; QuickBird, IKONOS, FORMOSAT-2, ALOS, SPOT-5, etc.
Output image format	8-bit, 3-band GeoTIFF only, easily compatible with GIS software, e.g., ArcGIS, MapInfo	8- or 16-bit; single- or multi-band; formats including BMP, ENVI, IDRISI, GeoTIFF
Output image type	True- or false-colour composite	True- or false-colour composite
Output image stretching	Both linear and non-linear image stretching options, accurate specification up to decimals	Both linear and non-linear image stretching options, rough specification up to integers
Processing speed	Very fast in processing GeoTIFF inputs	Slow in processing GeoTIFF inputs
Memory usage	Low	High
Fusion method	Local / global optimisations	Local / global optimisations
Spatial enhancement	None / detail / sharp	None / detail / sharp
Spectral enhancement	Option for greener vegetation	Option for greener vegetation
Feature enhancement	Vegetation to be highlighted in TOTAL green based on NDVI threshold
PAN/MS input pixel offset adjustment	X/Y directions
Targeted users	Intensive professional users - e.g., DigitalGlobe and GeoEye data vendors	Casual users
Potential system integration	Easy implementation and embedding with customer's image processing workflow and system	Limited batch processing facility

 

"An enormous amount of spatial data sets has been produced and explored already, but the future for a bright, sustained geospatial industry and location intelligence business totally depends on the high quality of input data."