Low-Cost 3D workshop

Brescia, Italy: 12-13 December 2024

8th International ISPRS Workshop

Low-Cost 3D - Sensors, Algorithms, Applications

Low-Cost 3D (LC3D) is a series of international ISPRS workshops on low-cost three-dimensional sensor systems and tools, ranging from low-cost acquisition devices like handheld scanning systems over inexpensive photogrammetric algorithms to processing software and applications. Initiated by Prof. Frank Neitzel from Technical University Berlin and Prof. Ralf Reulke from Humboldt University in Berlin, the workshops were initially organized and hosted in Berlin (2011, 2013, 2015). In 2017, the 5th workshop was hosted by the HafenCity University Hamburg; in 2019, the 6th edition took place at INSA Strasbourg, France, and in 2022, at the University of Würzburg, Germany.

The 8th edition of the ISPRS Low-Cost 3D workshop will be held at the University of Brescia in Brescia (Italy) from 12 to 13, 2024. 

The event will be in conjunction with the Optical 3D Metrology (O3DM) workshop. These two complementary ISPRS scientific events will provide delegates with a wide range of topics centred on 3D reconstruction and optical 3D measurement systems and applications. Participants could join both events with just one registration fee.

The main focus of the ISPRS LC3D workshop is to discuss new developments in low-cost 3D sensor technology, algorithms, and applications. 

Topics include:

 lC3D Invited SPEAKERS 

Anthony Pamart

UPR 2002 MAP CNRS, Models and Simulations for Architecture and Heritage, Marseilles. 

Toward synergies in open-source photogrammetry: expectations, blockers, and the MeshRoom X MicMac software prototype experiments. 

Since its emergence, free and open photogrammetry has followed a long and tortuous path. Nowadays and maybe more than even before the commercial solutions have spread their hegemony, fairly earned by performances, robustness and accessibility packed in user-friendly frameworks. At the age of open-science, the challenge still remains and requires the exploration of new models diffusing open-photogrammetry within the great family of Image-Based Modeling

techniques. What lessons can be learned from the early and pioneering experiences ? Can we transform and valorize academic born to be dead, because unsupported, initiatives to a stable, up to date and maintainable solutions ? Can we encompass the recently arisen techniques like RTI or NeRF ? Aside of these still open questions, the early stage and first experiments toward an open-source based interoperative prototype will be used as a discussion engine. 

KEYNOTE SPEAKERS

(shared with Optical3D Workshop workshop)

MARKUS GERKE

Technische Universität Braunschweig, Germany

Real-time visual localisation framework for satellite-based navigation

Reliable operation of global navigation satellite systems (GNSS) for positioning, navigation, and timing (PNT) is essential for applications, e.g., in communication and transport. Disruptions to GNSS threaten safety and security. The EU-funded Egeniouss project aims to enhance European GNSS services by creating an accurate, cost-effective cloud service. This service incorporates innovative multi-sensor navigation with visual localization to address GNSS challenges. Initially, it will feature an application programming interface for smartphones and drones, showcased in three case studies. Following successful validation, the service will expand for broader applications.


TILL SIEBERTH

Jade Hochschule, Oldenburg, Germany

State-of-the-art in forensic photogrammetry

Geosurveying has been used in police investigations for several years, mostly to survey traffic collisions. In recent years, optical 3D metrology has replaced typical accident surveying and has found its way into several other areas of forensic work, including forensic pathology. Specifically, low-cost sensors enable smaller police units to perform state-of-the-art incident documentation. However, optical 3D metrological data is not only acquired to document scenes and traces but also to perform advanced reconstructions of incidents. The aim is to provide deeper insight for the court and enable more adequate judgment. 


MICHAEL WEINMANN

TU Delft, The Netherlands

Traditional to AI-based 3D scene capture and modeling

Artificial intelligence (AI) has been demonstrated to offer breakthrough potential for numerous applications. This presentation will delve into how AI has revolutionized the field of 3D scene capture, creating more realistic and accurate 3D models. For this purpose, the presentation will review major developments in 3D scene capture from traditional methods to recent state-of-the-art techniques in terms of neural radiance fields (NeRF) and 3D Gaussian Splatting, as well as respective variants to further increase robustness and handle complex scenarios. Finally, the remaining limitations of current methods, as well as potential directions for future work, will be discussed.