FUSION

Data Fusion of Two Input Images


EnvironmentsPYTHON :: EASI
Quick linksDescription :: Parameters :: Parameter descriptions :: Details :: Examples :: Algorithm :: Related

Back to top

Description


FUSION creates an output RGB color image by fusing an input RGB color or pseudocolor image with an input black-and-white intensity image, using the IHS transform (Cylinder or Hexcone model) or the Brovey transform.
Back to top

Parameters


Name Type Caption Length Value range
FILE1 * String Input file name 1 1 - 192  
DBIC * Integer Input raster channels 1 - 3  
DBLUT Integer Lookup table segments 0 - 3  
DBPCT Integer Pseudocolor table segment 0 - 1  
FILE2 * String Input file name 2 1 - 192  
DBINT * Integer Input intensity channel 1 - 1 1 - 1024
DBINTLUT Integer Intensity lookup table segment 0 - 1  
FILO * String Output file name 1 - 192  
DBOC Integer Output raster channels 0 - 3  
RESAMPLE * String Resample mode 4 - 5 NEAR | BILIN | CUBIC
Default: NEAR
FUSMODEL * String Fusion model 6 - 8 CYLINDER | HEXCONE | BROVEY
Default: CYLINDER
MONITOR String Monitor mode 0 - 3 ON, OFF
Default: ON

* Required parameter
Back to top

Parameter descriptions

FILE1

Specifies the name of the image file containing three RGB color input channels or one pseudocolor input channel, used to define the color (hue and saturation) component of the fused output image.

The files specified by the FILE1 and FILE2 parameters must be in the same georeferencing units.

DBIC

Specifies input image channels for an RGB color image or a pseudocolor image (one channel with DBPCT specified).

Exactly one or three channels must be specified. If one channel is specified, then DBPCT must also be specified, and DBLUT is ignored.

If three channels are specified, then DBLUT can be (optionally) specified, and DBPCT is ignored.

Duplicate input channels are allowed.

16-bit or 32-bit values are linearly scaled to 8-bit values between 0 and 255 before applying a lookup or pseudocolor table.

DBLUT

Specifies three lookup table segments in the input file (FILE1) for enhancing three red, green, and blue input color channels (DBIC).

This parameter can be specified if DBIC specifies three channels. It is ignored if DBIC specifies only one channel.

If this parameter is not specified, three input channels are not enhanced.

DBPCT

Specifies the pseudocolor table segment in the input file (FILE1) for encoding one input channel (DBIC) into red, green, and blue values.

This parameter must be specified if DBIC specifies one channel. It is ignored if DBIC specifies three channels.

FILE2

Specifies the name of the input file that contains the input intensity image for data fusion.

The files specified by FILE1 and FILE2 must be in the same georeferencing units.

DBINT

Specifies one input black-and-white intensity image in the file specified by FILE2. DBINT must not be the same as DBOC.

Supported values are between 1 and 1024. 16-bit or 32-bit values are linearly scaled to 8-bit values between 0 and 255 before applying the lookup table.

DBINTLUT

Specifies one lookup table segment in the file specified by FILE2 for enhancing the intensity input channel (DBINT).

If not specified, the intensity channel is not enhanced.

FILO

Specifies the name of the output file that contains three output color channels (DBOC) resulting from the data fusion.

FILE1, FILE2, and FILO can specify the same files. If the specified output file does not exist, a new file is created using the georeferencing bounds and minimum pixel resolution from FILE1 and FILE2.

DBOC

Specifies three output RGB channels in the file specified by FILO to store the results of data fusion. The specified DBIC channels in FILE1 define the color, and the specified DBINT channel in FILE2 defines the intensity for the fused output color image.

DBOC must be specified if the file specified by FILO exists; if it does not exist, DBOC defaults to 1, 2, and 3.

DBOC must not be the same as DBINT. Duplicate output channels are NOT allowed.

Results are always 8-bit values between 0 and 255.

RESAMPLE

Specifies the type of resampling method used on both input images.

Resampling extracts and interpolates the gray levels from the original pixel locations to corrected locations.

Three resampling modes are supported:

The default value is NEAR (NEAREST).

FUSMODEL

Specifies the type of fusion model to use. Valid values are:

Refer to the 'Details' and 'Algorithm' sections for more information about these models.

MONITOR

The program progress can be monitored by printing the percentage of processing completed. A system parameter, MONITOR, controls this activity.

Available options are:

Back to top

Details

FUSION creates an output RGB color image by fusing an input RGB color or pseudocolor image with an input black-and-white intensity image, using the IHS transform (Cylinder or Hexcone model) or the Brovey transform. Data fusion combines the color component of one input image with the intensity component of another image, to create a new fused color output image.

FUSION performs data fusion of a Red-Green-Blue color image (DBIC in FILE1) or of an input pseudocolor image (DBIC and DBPCT in FILE1) with a black-and-white intensity image (DBINT in FILE2), using one of three different fusion models: Cylinder, Hexcone, or Brovey (as specified by the FUSMODEL parameter).

The Hexcone model is used by many commercial image processing software products. The Hexcone model runs about 15 percent faster than the Cylinder model.

The Brovey transform is a highly effective transform that generates a better looking image than the normal RGB image for many types of data, in particular for combining Landsat TM and SPOT Pan imagery. The Brovey transform is a formula-based process based on the band to display in a given color, the sum of all the color layers, and the intensity layer (DBINT).

If FILE1 and FILO are the same and FILE2 is different, then the black-white intensity image (DBINT) is resampled using the specified resampling method (RESAMPLE).

If FILE2 and FILO are the same and FILE1 is different, then the RGB input image or pseudocolor input image is resampled using the specified resampling method (RESAMPLE).

If the output file specified by FILO does not exist, the new file is created with the georeferencing bounds of FILE1 and FILE2 and with the higher resolution of the two of them (FILE1 or FILE2).

The lower resolution image is resampled using the specified resampling method (RESAMPLE).

If the output file specified by FILO exists and it is different from those specified by FILE1 and FILE2, then both FILE1 and FILE2 files are resampled to the FILO resolution using the specified resampling method (RESAMPLE).

FILE1 specifies the file containing the image data. The three input channels specified in DBIC represent the red, green, and blue channels, in that order. If one input channel is specified, the PCT segment must be specified. The three output channels specified in DBOC are used for the output of red, green, and blue , in that order.

FILE1, FILE2, and FILO (if the specified file exists) must be in the same georeferencing units. The number of input channels (DBIC) must be one or three.

Back to top

Examples

This example fuses a 30-meter resolution Thematic Mapper (TM) color image in irvine.pix with a 10-meter resolution panchromatic SPOT image in eltoro.pix. The result is a 10-meter resolution color image.

Step 1: For better results, enhance the RGB and intensity images first.

EASI>FILE="irvine.pix"
EASI>DBIC=1,2,3
EASI>DBLUT=2,3,4
EASI>DBOC=1,2,3
EASI>MASK=
EASI>RUN LUT
EASI>FILE="eltoro.pix"
EASI>DBIC=1
EASI>DBLUT=2
EASI>DBOC=1
EASI>MASK=
EASI>RUN LUT

Step 2: Add three 8-bit channels to eltoro.pix to store the result.

EASI>FILE="eltoro.pix"
EASI>PCIOP="ADD"
EASI>PCIVAL=3
EASI>RUN PCIMOD

Step 3: Use FUSION to fuse the data. Both the Cylinder model and the Hexcone model produce good results, but the Cylinder model has fewer false artifacts.

EASI>FILE1="irvine.pix"
EASI>DBIC= 3,2,1
EASI>DBPCT= 
EASI>FILE2= "eltoro.pix"
EASI>DBINT= 1
EASI>FILO= "eltoro.pix"
EASI>DBOC= 2,3,4
EASI>RESAMPLE ="CUBIC"
EASI>FUSMODEL="CYLINDER"
EASI>BACKVAL=
EASI>RUN FUSION

This example fuses the USGS Land Use/Land Cover theme map on channel 6 of irvine.pix with a 10-meter resolution panchromatic SPOT image in eltoro.pix. The result is a 10-meter resolution SPOT image with the land use classes overlaid in translucent colors.

Step 1: Add three 8-bit channels to eltoro.pix to store the result.

EASI>FILE="eltoro.pix"
EASI>PCIOP="ADD"
EASI>PCIVAL=3
EASI>RUN PCIMOD

Step 2: Enhance the input intensity image in eltoro.pix. Otherwise, the pseudocolor image colors overpower the intensity image.

EASI>FILE="eltoro.pix"
EASI>DBIC=1
EASI>DBLUT=2
EASI>DBOC=1
EASI>MASK=
EASI>RUN LUT

Step 3: Use FUSION to fuse the data. Nearest neighbor resampling must be specified, since the input image is a theme map. In this case, the Hexcone model produces better results than the Cylinder model.

EASI>FILE1="irvine.pix"
EASI>DBIC=6
EASI>DBLUT=
EASI>DBPCT=7
EASI>FILE2="eltoro.pix"
EASI>DBINT=1
EASI>FILO="eltoro.pix"
EASI>DBOC=2,3,4
EASI>RESAMPLE="NEAR"
EASI>FUSMODEL="HEXCONE"
EASI>RUN FUSION
Back to top

Algorithm

The equations used to convert Red, Green, and Blue color values to Intensity, Hue, and Saturation color values for both the Cylinder and the Hexcone IHS models are described in the 'Algorithm' section of the RGB and IHS documentation.

The Brovey transform is a formula-based process that works by dividing the band to display in a given color by the sum of all the color layers (for example, Red, Green, and Blue) and then multiplying by the intensity layer (for example, SPOT Panchromatic layer).

R = Red, G = Green, B = Blue , I = Intensity (Panchromatic)

Red layer:

(R / (R + G + B)) * I

Green layer:

(G / (R + G + B)) * I

Blue layer:

(B / (R + G + B)) * I

© PCI Geomatics Enterprises, Inc.®, 2026. All rights reserved.