RPC Bias-correction

To correct RPCs for biases, ensure the project contains images with corresponding RPCs and 2D points, and 3D ground control points in geographic coordinates. All these elements should be set active (highlighted green). Check your sigma values before commencing any processing.

1. RPC bundle adjustment: Shift

Adjustment > RPC Bundle > Set bias correction parameters > Shift

You may notice that by setting bias correction parameters to Shift, the sigma values of the bias parameters A0 and B0 have changed in the RPC files.

Adjustment > RPC Bundle > Run Bundle

A dialog box should notify you of Adjustment success or failure. On success, a XYZ node labelled ‘RPCBundle’ will be added in the Items window. Double-Click on the RPCBundle node will display the adjustment results.

2. Check your results

Image residuals can be viewed on any image used in the bundle adjustment. They appear as pink vectors originating from the point’s origin. If they are not visible, they may be too small. Pressing the + and - keys on the keyboard will scale the residuals. The scale factor can be found on the bottom edge of the display.

Image residuals are displayed on the image as pink vectors

To compare the Adjustment calculated positions to the ground control points, Right-Click on the RPCBundle node > Difference by label. Select GCPs_WGS84Geogrpahic.csv and select the units for coordinate differences (metres). A XYZdiff node will appear in the Items menu, storing the calculated differences between the adjustment results and the ground control results for each point and a corresponding RMS value. With this dataset, the RMS should be to the order of half a pixel (0.35 m).

3. Regenerate and Export RPCs

Adjustment > Regenerate bias-corrected RPCs

Adjustment > Export updated RPCs > QuickBird format

This allows easy import into another Barista project or into another software package.

Now the project has accurate sensor orientation, we can generate accurate 3D products, including monoplotting, orthoimage generation and pansharpening.

To Sample Affine Model

A. Project setup
B. Image measurement
D. Monoplotting
E. Orthoimage generation
F. Pansharpening

Back to Sample Project Menu