Realtime projective multi-texturing of pointclouds and meshes for a realistic street-view web navigation

Alexandre Devaux (1) , Mathieu Brédif (2)

1 LaSTIG - Laboratoire des Sciences et Technologies de l'Information Géographique
2 MATIS - Méthodes d'Analyses pour le Traitement d'Images et la Stéréorestitution

Street-view web applications have now gained widespread popularity. Targeting the general public, they offer ease of use, but while they allow efficient navigation from a pedestrian level, the immersive quality of such renderings is still low. The user is usually stuck at specific positions and transitions bring out artefacts, in particular parallax and aliasing. We propose a method to enhance the realism of street view navigation systems using a hybrid rendering based on realtime projective texturing on meshes and pointclouds with occlusion handling, requiring extremely minimized pre-processing steps allowing fast data update, progressive streaming (mesh-based approximation, with point cloud details) and unaltered raw data precise visualization.

Mots clés: Image Based Rendering; Projective Texturing; Street- view; WebGL; Point Based Rendering

Domaines: Ingénierie de l'environnement

Alexandre Devaux, Mathieu Brédif. Realtime projective multi-texturing of pointclouds and meshes for a realistic street-view web navigation. 21st International Conference on Web3D Technology, Jul 2016, Anaheim, United States. pp.105-108, ⟨10.1145/2945292.2945311⟩. ⟨hal-02290370⟩