Take me to pcigeomatics.com

PCI Geomatics Help Center

How can we help you today?

ALAGAC ZY3 PanSharpening Workflow

PCI Geomatics -

Summary

The following document describes how to process ZY3 satellite imagery to create a Pan-sharpened Image. The workflow is comprised of 6 main procedures:

  1. Stitch Panchromatic and Multispectral Images (Optional)
  2. Photogrammetric Processing of Panchromatic Imagery
    • GCP collection – using Google Earth
    • Automatic Tie Point collection
    • Orthorectify Panchromatic Images
  3. Mosaic Orthorectified Panchromatic Images
  4. Multispectral GCP collection
    • GCP collection - using Mosaic Orthorectified Panchromatic image as reference
    • Orthorectify Multispectral Images
  5. Mosaic Orthorectified Multispectral Images
  6. Pan-sharpen imagery

The result of the following workflow is a seamless Pan-sharpened ZY3 mosaic.

Stitch Panchromatic and Multispectral Images

This step is optional.  Stitching can only be performed using scenes acquired on the same path and same day. This process eliminates mosaicking and generates a long continuous strip which reduces the number of GCPs required.

  1. Create a new OrthoEngine project as seen below.

image001.png

  1. When the projection window appears click Cancel.
  2. Open the Stitch Image Tiles with RPC tool located in the tools tab.
  3. Select OTHERS in Satellite.
  4. Input each xml file from the Panchromatic image folders for the images in the same flight path and in sequential order

image003.png

  1. For image folder zy301a_nad_032074_016116_20171018110726_01_sec_0001_1710191646 the user only needs to import the xml file as there are no other images in the flight path. Select the Data Input Processing step and click on the Open Images button image005.png, add xml file.
  2. When completed, the Project Overview window should be similar to the image below.

image007.png

  1. Save and close the project
  2. Start a new project and follow the same steps for the multispectral imagery.

 

Photogrammetric Processing of Panchromatic Images

The photogrammetric processing of the stitched panchromatic images to produce orthorectified images. This procedure is conducted in OrthoEngine. Google Earth will be used as the reference image for the GCP collection. A GCP should be captured in each of the image’s corners and inside the image (such as near the centre of the image).

  1. Open the Panchromatic project, select GCP/TP Collection.
  2. Open the manual GCP tool image009.png
  3. In the GCP tool specify the DEM and check Compute Model
  4. To collect coordinates from Google Earth:
    1. Open working image in Focus
    2. Locate potential GCP collection sites and place cursor on location
    3. Select the File tab and Export to Google Earth
    4. Find the same location in Google Earth 
    5. Select Add Placemark image011.png and choose the location to obtain the coordiantes As seen in the image below

image012.png

    1. Enter the coordinate information into the GCP collection window and select Extract Elevation.
    2. Although the cursor will automatically orient to the input location in the GCP image viewer, the user will should adjust the placement to ensure it is on the same spot.
  1. As GCPs are collected the Residual values will be computed for accuracy. GCP’s with residuals greater than 1 pixel need to be reassessed and checked. This can be done by selecting on the Point ID table in the GCP collection window which will display the location of the GCP point.
  2. Once a sufficient number of GCP’s have been collected with low residual values close the GCP collection window and save your project.
  3. Open the automatic tie point collection tool image014.png
  4. In the TP Collection tool, specify the following parameters:
    1. Distribution Pattern: Overlap
    2. TPs per area: 100
    3. Sample Source Method: SUSAN

image016.png

  1. Click collect tie points.
  2. Once Automatic Tie Point Collection has been completed close the window and open the residual report image018.png. If the values are red then click the compute model button.
  3. If TP’s collected have residuals above 1 pixel then remove them from the model.

image020.png

  1. Once the residual values are below 1 then save the project and close the residual report.
  2. In the Processing step menu, select Ortho Generation.
  3. Open the Ortho tool image022.png
  4. In the Ortho tool:
    1. Select all the images to process
    2. Specify an output folder
    3. Select a DEM
    4. Click Generate Orthos
  5. Save the project

 

Mosaic Ortho Panchromatic Images

Create a seamless mosaic of the Ortho Panchromatic images.

  1. Choose the Mosaic Processing step and select Define the Mosaic output
    1. Create the Mosaic file
  2. Select Automatic mosaicking step
    1. Select Generate Mosaic
  3. Save and close project.

 

Photogrammetric Processing of Multispectral Images

The photogrammetric processing of the stitched Multispectral images to produce ortho images.

  1. Open the Multispectral Project created in the Stitching step.
  2. In the processing step menu, select GCP/TP Collection.
  3. Open the automatic GCP tool image024.png
  4. In the GCP tool specify the following:
    1. Specify the Mosaic panchromatic image
    2. Specify the DEM
    3. Check Stereo GCP
    4. Keep all settings as default
    5. Click Match GCPs
    6. Click Add GCPs to Project
  5. Close the automatic GCP tool and open the residual report.
  6. As the Multispectral mosaic contain a lower resolution than the Panchromatic any error inherent in the GCPs will be 3 times greater. As a result, all GCPs that contain residual values greater than 0.3 need to be removed.
  7. Now is a good time to save the project.
  8. In the Processing step menu, select Ortho Generation.
  9. Open the Ortho tool image026.jpg
  10. In the Ortho tool:
    1. Select all the images to process
    2. Specify an output folder
    3. Select a DEM
    4. Click Generate Orthos
  11. Save the project

 

Mosaic Orthorectified Multispectral Images

Create a seamless mosaic of the Ortho Multispectral images.

  1. Choose the Mosaic Processing step and select Define the Mosaic output
    1. Create the Mosaic file
  2. Select Automatic mosaicking step
    1. Select Generate Mosaic
  3. Save and close project.

 

Pan-sharpen Mosaic Images

The next step is the pan-sharpening of the MS and Pan mosaics.

  1. In Focus, go to the Tools menu, and select Algorithm Librarian to open the Algorithm Librarian.
  2. Click on Find to open the search panel.
  3. In the search panel, type PANSHARP2 and click on find.
  4. Double-click on the PANSHARP2 tool to open.
  5. Browse and open both Panchromatic and Multispectral mosaics.
  6. In the multispectral image input, specify all bands.
  7. In the reference image input, specify all bands.
  8. In the panchromatic input, specify the panchromatic band.
  9. In the output, specify an output file.
  10. Click Run.
  11. To test the accuracy of the Pan Sharpened image, zoom to the town located in the red square.

image027.png

  1. Viewing the buildings at 1:1 Resolution.
  2. If the roof colors are bleeding as seen in the image below than the GCPs of the Multispectral imagery need to be reassessed and refined.

image029.png

  1. Correct PanSharp results.

image031.png

 

Have more questions? Submit a request

Comments

Powered by Zendesk