2017 TIM workshop

Magellium attends the DGA TIM workshop to show its recent works related with semantic segmentation. These works include automatic building footprint extraction, automatic land cover classification from simulated Sentinel-2 images, ship detection (on behalf an undisclosed customer), handwritten digits recognition from 18th-19th century maps and on-going works on cloud and vehicle detection.

Transcript

1. Applications du deep learning à l’analyse d’image « De la donnée satellite au contenu sémantique » Notre expertise Multidisciplinaire  Géomatique  Télédétection  3D, Computer Vision  Calcul numérique Système  Interfaces métier  Déploiement Cloud Notre offre Expertise métier, conseil Intégration dans votre SI Réseau clé en main Automatic Building Footprint Extraction Aerial image Pixel segmentation Vector footprint Ship Detection (on behalf of ) Automatic Land Cover Classification (RGB, 10m GSD from Pléiades)