‪Roger Marí‬

I am a computer vision research engineer with a PhD and a strong background in image processing and deep learning. My thesis focused on remote sensing and 3D vision tasks (reconstruction, calibration, co-registration, change detection, etc.). My latest works focus on neural rendering techniques applied to remote sensing images, the digitization of cultural heritage and the exploration of generative AI.

I am from Barcelona (1995), former student of Universitat Pompeu Fabra. In Barcelona I completed with honors a BSc in Audiovisual Systems Engineering and then a specialized MSc in Computer Vision. I moved to Paris in October 2018 to pursue my PhD under the supervision of Gabriele Facciolo, at the Centre Borelli (ENS Paris-Saclay). I defended my thesis Applications of multi-image remote sensing in December 2022. In January 2024 I joined Eurecat where I currently contribute to the coordination and development of AI and computer vision projects in Catalonia and Europe.

Email  /  CV  /  Google Scholar  /  ResearchGate  /  LinkedIn  /  GitHub

profile photo
Research

I am interested in computer vision, machine learning, optimization, and image processing. I am particularly passionate about 3D models and the whole process behind their acquisition, generation and assessment.

Multi-Date Earth Observation NeRF: The Detail Is in the Shadows
Roger Marí‬, Gabriele Facciolo, Thibaud Ehret
CVPR Workshops, 2023
project page / paper / poster

We present EO-NeRF, that reveals scene geometry from multi-date satellite images with an unprecedented level of detail. We propose a geometrically consistent shadow model and a radiometric decomposition of the scene more adapted to pansharpened satellite images.

blind-date Disparity Estimation Networks for Aerial and High-Resolution Satellite Images: A Review
Roger Marí‬, Thibaud Ehret, Gabriele Facciolo
IPOL, 2022
paper / demo

We evaluate the performance of the deep learning architectures PSM (CVPR 2018) and HSM (CVPR 2019) for disparity estimation on multiple pairs of high-resolution satellite images.

Regularization of NeRFs Using Differential Geometry
Thibaud Ehret, Roger Marí‬, Gabriele Facciolo
Preprint, 2022
paper

We propose a generic regularization framework for NeRF based on differential geometry that outperforms previous state-of-the-art methods with only three input views. We compare our approach with RegNeRF (CVPR 2022).

blind-date Sat-NeRF: Learning Multi-View Satellite Photogrammetry With Transient Objects and Shadow Modeling Using RPC Cameras
Roger Marí‬, Gabriele Facciolo, Thibaud Ehret
CVPR Workshops, 2022
project page / paper / code / poster

Sat-NeRF is the first work in neural rendering for multi-date satellite images to demonstrate quantitatively convincing results in terms of surface reconstruction.

L1B+: A Perfect Sensor Localization Model for Simple Satellite Stereo Reconstruction from Push-Frame Image Strips
Roger Marí‬, Thibaud Ehret, ‪Jérémy Anger, Carlo de Franchis, Gabriele Facciolo
ISPRS Annals, 2022
paper / poster

We emulate a perfect sensor to generate a single image from a fragmented push-frame strip. The resulting product simplifies large-scale 3D modeling from push-frame imagery.

blind-date A Generic Bundle Adjustment Methodology for Indirect RPC Model Refinement of Satellite Imagery
Roger Marí‬, Carlo de Franchis, Enric Meinhardt-Llopis, ‪Jérémy Anger, Gabriele Facciolo
IPOL, 2021
paper / demo / code

We propose a generic bundle adjustment method for multi-view stereo pipelines for satellite images. The RPC camera models of the input views are refined with a rotation that compensates localization errors related to the attitude angles encoding the satellite orientation.

Automatic Stockpile Volume Monitoring Using Multi-View Stereo from SkySat Imagery
Roger Marí‬, Carlo de Franchis, Enric Meinhardt-Llopis, ‪Gabriele Facciolo
IGARSS, 2021
paper

The RPC camera models of a time series of SkySat acquisitions are refined and used to compute a surface model for each date, which is used to measure the stockpile volume.

blind-date Robust Rational Polynomial Camera Modelling for SAR and Pushbroom Imaging
Roland Akiki, Roger Marí‬, Carlo de Franchis, Jean-Michel Morel, ‪Gabriele Facciolo
IGARSS, 2021
paper / code

We describe a terrain-independent algorithm to accurately derive the RPC camera model linking a set of 3D-2D point correspondences based on a regularized least squares fit.

blind-date To Bundle Adjust or Not: A Comparison of Relative Geolocation Correction Strategies for Satellite Multi-View Stereo
Roger Marí‬, Carlo de Franchis, Enric Meinhardt-Llopis, ‪Gabriele Facciolo
ICCV Workshops, 2019
project page / paper / poster

This work investigates and compares different relative geolocation correction techniques for multi-view stereo pipelines for satellite images. We assess the impact on the output geometry.

Deep Single Image Camera Calibration with Radial Distortion
Manuel López-Antequera, Roger Marí‬, Pau Gargallo, Yubin Kuang, Javier Gonzalez-Jimenez, Gloria Haro
CVPR, 2019
paper / supp

We present a deep learning method to predict extrinsic (tilt and roll) and intrinsic (focal length and radial distortion) parameters from a single image. We use a parameterization that is better suited for learning than directly predicting the camera parameters.


Design and source code from Jon Barron's website.