Skip to content

Centre Borelli moduls for calibrating RPCs with L-curve regularization

License

Notifications You must be signed in to change notification settings

Hydrosat/rpcfit

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

rpcfit

rpcfit is a python module that enables the robust fitting of rational polynomial camera models.

The algorithm is described in the IGARSS 2021 paper:

Robust Rational Polynomial Camera Modelling for SAR and Pushbroom Imaging

doi: 10.1109/igarss47720.2021.9554583

Preprint available at arxiv and hal

Citation:

@inproceedings{akiki2021robust,
  title={Robust rational polynomial camera modelling for SAR and pushbroom imaging},
  author={Akiki, Roland and Mar{\'\i}, Roger and De Franchis, Carlo and Morel, Jean-Michel and Facciolo, Gabriele},
  booktitle={2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS},
  pages={7908--7911},
  year={2021},
  organization={IEEE}
}

Installation

To reproduce the environment used for the paper, checkout the tag "paper", then install requirements :

pip install -r requirements.txt

Content

To test the installation: python usage.py

The package contains:

  • gridata.py: Has the necessary functions to construct the 3D+2D point grid correspondence (CoNtrol Point & ChecK Point). Needs a projection function or a localization function from a physical sensor model. A physical sensor model for the Sentinel-1 satellite may be made available in the future. Otherwise, consider using snappy for Sentinel-1.

  • Lcurve.py: Has the necessary function to do the Lcurve criterion with the standard Tikhonov, or with a set of discrete points by using the spline curve interpolation.

  • rpc_fit: Has the necessary functions to fit the rpc on the constructed grids. use calibrate_rpc. Other functions are internal helper funcs.

The data folder contains datasets for the test in usage.py:

  • 2d refers to image coordinates x, y
  • 3d refers to lon, lat, height coordinates
  • train refers to the training set
  • test refers to the test set

About

Centre Borelli moduls for calibrating RPCs with L-curve regularization

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 100.0%