Diffusion anisotrope et inpainting d'orthophotographies LiDAR mobile

Abstract : This paper proposes a pipeline for the joint generation of reflectance and height orthophotography at very high resolution (1cm) from a mobile laser scan. The ground projection of the LiDAR point cloud over a 2D pixel grid provides a sparse and noisy image of both channels. We offer to enhance those images by generating a dense image using anisotropic diffusion and TV-L 1 denoising. Moreover , we adapt an inpainting method in order to fill large occlusions by taking advantage of the dense images and an assumption of the alignment between structures to inpaint.
Document type :
Conference papers
Liste complète des métadonnées

Cited literature [6 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01316487
Contributor : Pierre Biasutti <>
Submitted on : Tuesday, May 17, 2016 - 10:58:09 AM
Last modification on : Wednesday, October 10, 2018 - 11:30:06 AM
Document(s) archivé(s) le : Friday, August 19, 2016 - 4:20:35 PM

File

biasutti2016orthophoto.pdf
Files produced by the author(s)

Licence


Distributed under a Creative Commons Attribution 4.0 International License

Identifiers

  • HAL Id : hal-01316487, version 1

Collections

Citation

P Biasutti, Jean-François Aujol, Mathieu Brédif, Aurélie Bugeau. Diffusion anisotrope et inpainting d'orthophotographies LiDAR mobile. 20ème Congrès national sur la Reconnaissance des Formes et l'Intelligence Artificielle (RFIA 2016), Jun 2016, Clermont-Ferrand, France. ⟨hal-01316487⟩

Share

Metrics

Record views

201

Files downloads

132