Image Fusion

Instructions for Use

Image fusion is a technique that combines and transforms multiple images from different times, sensor systems, resolutions, and spectral bands through various methods to generate new images. For example, fusing a panchromatic orthoimage with a multispectral orthoimage can produce high-quality imagery with both high spatial and spectral resolution. This technology effectively combines the advantages of various types of imagery, providing richer information support for subsequent analysis and applications.

SuperMap iDesktopX supports this feature starting from version 11i(2023).

Figure: Multispectral Before Fusion Figure: Panchromatic Before Fusion Figure: Multispectral After Fusion

Parameter Description

Parameter Name Parameter Interpretation Parameter Type
Data Type

This parameter is available only when the function is launched from the Imagery tab -> DOM Data Production group -> Image Fusion.

Select the source of the input image data, with two options: Image Files and Mosaic Dataset. The default is Image Files. Use the Image File List to manage image files with functions like Add, Select All, Select Reverse, Delete, and Settings.

  • Add: Provides four ways to input image files: Add File, Add Folder, Add List, and Add Mosaic Data. When selecting Add Folder, all image files in the folder (including subfolders) are automatically retrieved.
  • Select All: Selects all records in the list.
  • Select Reverse: Inverts the selection of all records in the list.
  • Delete: Deletes the selected records from the list.
  • Settings: When the input panchromatic and multispectral images are not fully paired, use the Settings button to manage pairing rules to complete the pairing for all images.

 
Panchromatic Data Source

This parameter is available only when the function is launched from the Imagery tab -> DOM Data Production group -> Image Fusion.

After selecting Mosaic Dataset in Data Type, specify the data source where the panchromatic image data is located.

DataSource
Panchromatic Data After selecting Mosaic Dataset in Data Type, specify the mosaic dataset where the panchromatic image data is located. If it is part of a workflow production, it is display-only and cannot be edited. DatasetMosaic
Multispectral Data Source

This parameter is available only when the function is launched from the Imagery tab -> DOM Data Production group -> Image Fusion.

After selecting Mosaic Dataset in Data Type, specify the data source where the multispectral image data is located.

DataSource
Multispectral Data After selecting Mosaic Dataset in Data Type, specify the mosaic dataset where the multispectral image data is located. If it is part of a workflow production, it is display-only and cannot be edited. DatasetMosaic
Fusion Method

Specifies the image fusion method. The default method is Gram-Schmidt.

  • PanSharp: An image fusion method based on the gray-scale relationship of least squares. It establishes a linear regression model between multispectral and panchromatic images, adjusts the gray values of each band in the multispectral image to match the radiation characteristics of the panchromatic image, thereby reducing color deviation after fusion. Suitable for remote sensing applications requiring high spectral accuracy, such as agricultural monitoring and environmental assessment.
  • Gram-Schmidt: An improved method based on Gram-Schmidt orthogonal transformation. It adaptively adjusts the linear relationship between each band of the multispectral image and the panchromatic image to generate a brightness component that better meets practical needs. This method features high spectral fidelity and good spatial detail preservation.
  • CNBROVEY: CNBROVEY (short for Color Normalized Brovey) is an improved Brovey transform-based image fusion method. It optimizes the fusion process of multispectral and panchromatic images by introducing a color normalization strategy. It enhances spatial details while preserving the original spectral characteristics as much as possible, suitable for scenarios balancing spectrum and detail.
  • HSV: HSV (short for Hue, Saturation, Value) is an image fusion method based on color space. It converts the multispectral image from RGB color space to HSV space, then uses the high spatial resolution of the panchromatic image to replace or enhance the value component, thereby improving image details while preserving hue and saturation. Suitable for RGB band fusion; less effective for non-visible light bands (e.g., near-infrared). Supports only 8-bit image data.
  • PCA: PCA (short for Principal Component Analysis) is an image fusion method based on statistical feature extraction. It converts the multispectral image into the principal component space and uses the panchromatic image to replace the first principal component (PC1), achieving effective fusion of spatial details and spectral information. This method enhances the spatial details of multispectral images while retaining most spectral information, suitable for remote sensing applications requiring high spatial resolution. However, it may risk spectral distortion, and spectral fidelity is generally moderate.
  • NND: NND (short for Non-Negative Decomposition) is an image fusion method based on non-negative matrix factorization (NMF). It jointly decomposes multi-source images (e.g., multispectral, panchromatic, infrared) to extract shared features (e.g., spatial structure) and independent features (e.g., spectrum, thermal radiation), and performs fusion based on this. Suitable for fusion between multi-modal data, but computation time is relatively longer.
FusionMethod
Predefined Fusion Weights Check this box to use predefined weights for fusion, but it is only effective for the PanSharp and Gram-Schmidt methods. If unchecked, adaptive weights are used for fusion. Boolean
Sharpening Factor Used to adjust the clarity of the fused image. The default is 1. Double
Resampling Method Sets the resampling method after image fusion. Options include Nearest, Bilinear, and Cubic. The default is Bilinear. For more method details, refer to Resampling Method. ResampleMethod
Export Directory

Sets the storage path for the result image.

If executing a production workflow, use the Process parameter in Set Image Path or Add Orthophoto, and this field is not editable here.

File
Output Format

Sets the storage type for the result image. Options include Not Set, Tiff/GeoTIFF, Erdas Image, and PCIDSK. The default is Tiff/GeoTIFF. Selecting Not Set means using the original storage type of the image for output.

If executing a production workflow, use the Output Format parameter in Set Image Path or Add Orthophoto, and this field is not editable here.

FileType
Compression Method

Sets the compression method for the output image. Available only when the output format is Tiff/GeoTIFF. The default is NONE.

If executing a production workflow, use the Compression Method parameter in Set Image Path or Add Orthophoto, and this field is not editable here.

CompressMethod
Add Results to Map Window

This parameter is available only when the function is launched from the Imagery tab -> DOM Data Production group -> Image Fusion.

Checked by default, meaning the result data will be added to the map window for display after successful execution.

Boolean
Create Image Pyramid Checked by default. Builds pyramids for the image data based on the specified resampling method to improve image zoom display rendering and performance. Boolean
Resampling Method Available after checking Create Image Pyramid. The default is Average. For more method details, refer to Resampling Method. PyramidResampleType
Asynchronously Create Statistics During image browsing, the layer will be stretched based on statistical information. If the image lacks statistical information, it will be displayed using the default stretch method, and 8-bit data will not be stretched. Checked by default, meaning statistics will be created for the image in the background without affecting workflow execution. If unchecked, statistics will be created during tool execution, which may take additional time. Boolean

Output Result

  • Outputs image data combining features of panchromatic and multispectral images to a local folder.
  • Automatically creates an image fusion mosaic dataset in the OrthoData data source based on the output image data.

Related Topics

Generate Orthophotos

Multispectral and Panchromatic Registration

True Color Output

Image Color Balancing

Image Mosaic