CBL - Campus del Baix Llobregat

Projecte llegit

Títol: Web-Based Real-Time Point Cloud Visualization in LiDAR-Equipped UAS

Estudiants que han llegit aquest projecte:

Director/a: ROYO CHIC, PABLO

Departament: DAC

Títol: Web-Based Real-Time Point Cloud Visualization in LiDAR-Equipped UAS

Data inici oferta: 27-01-2025     Data finalització oferta: 27-09-2025


Estudis d'assignació del projecte:
    DG ENG AERO/SIS TEL
Tipus: Individual
 
Lloc de realització: EETAC
 
Paraules clau:
LiDAR, UAS, 3D, Point Cloud
 
Descripció del contingut i pla d'activitats:
El estudiante deberá desarrollar un software para capturar datos de un LiDAR 3D instalado en un dron y visualizarlo en tierra en tiempo real.

Tareas a realizar:
1. Captura de datos LiDAR a bordo del dron con un computador
2. Envío de datos por LTE o radioenlace a tierra de forma eficiente.
3. Desarrollo de un software para visualizar la nube de puntos enviada desde el dron.
 
Overview (resum en anglès):
In recent years, advancements in LiDAR sensing have enabled increasingly sophisticated applications in surveying, mapping, and environmental monitoring. However, most existing point cloud visualization tools are designed for offline processing and cannot meet the latency and adaptability requirements of live data acquisition scenarios. This thesis addresses this challenge within the framework of the GAMMA-DRON project, whose objective is to provide to emergency first-responders real-time situational awareness during radiological spectrometry campaigns. The work focuses on the development of a web-based system capable of visualizing LiDAR point clouds in real time, integrating radiological measurements directly into the 3D scene for immediate spatial correlation. The developed system employs DJI's Matrice 350 RTK and Zenmuse L2 LiDAR sensor to generate high-density 3D data. Recent state-of-the-art research solutions for spatial indexing have been leveraged, notably the Lidarserv server, which supports efficient incremental point cloud updates and low-latency querying. On top of this backend, a custom viewer has been implemented entirely from scratch for the web environment. The viewer supports progressive rendering of massive point clouds, dynamic level-of-detail management, GPU-accelerated shading techniques, and precision-preserving coordinate transformations, allowing smooth interaction even with billions of points. Radiological dose rate measurements are processed and integrated seamlessly into the visualization pipeline. Interpolation algorithms generate continuous spatial fields from discrete measurements, enabling the user to intuitively identify and assess radiation patterns in direct relation to physical structures and terrain features. The resulting system bridges a critical gap in the state of the art, delivering a platform that combines the responsiveness of real-time data streaming with the accessibility of browser-based deployment, and establishing a technical foundation for future expansion into large-scale operational and research applications.

© CBLTIC Campus del Baix Llobregat - UPC