About This Project

Reality modeling algorithms allow the automatic generation of 3D models from photographs. Traditionally, algorithms are developed for “pinhole images”, which are the sensors in smartphones and cameras.

Nowadays, new 360 degree cameras are available. I will develop a novel method for accurate 3D modeling with cheap 360 cameras, which capture the entire scene around the photographer in a single shot. Results will be shared on a website to give people the opportunity to create virtual 3D models.

Ask the Scientists

What is the context of this research?

Standard methods for image-based 3D modeling are based on traditional cameras with a limited field of view, which require a large number of photographs acquired following specific rules. The other solution is laser scanning, but scanning sensors are very expensive (> 50,000 $).

My idea is to work with affordable 360 cameras. The 360 camera can be pointed to any direction and the very large field of view reduces the number of photographs. The entire scene around the camera is acquired in a single shot. Starting from some images (at least 2) acquired from different points, a 3D model can be generated in an automatic way. I am currently working on this topic and more results can be found in my paper.

What is the significance of this project?

I teach students advanced 3D modeling techniques in my daily activity at the university. I know the importance of accurate 3D measurements in several practical applications also in different fields: architecture, engineering, geology, archeology, criminology, etc.

Recording shape and size is a fundamental requirement in many disciplines!

The challenge is to provide photographers (equipped with an affordable 360 camera!) a simple and accurate solution for rapid mapping, with also a direct connection to virtual reality. 3 low-cost cameras will be tested to validate the method for 3D modeling. I will provide specific software that does not require expensive workstation and simple guidelines to replicate the process

What are the goals of the project?

The aim is to provide tutorials and datasets on a website. First, I will provide simple reports about the results achieved in terms of costs and metric accuracy. Tutorials will contain guidelines on the image acquisition procedure and processing. Results will be published online at two levels:

(1) a basic level for people interested in the proposed method for their surveying application

(2) an advanced level for surveying specialists

Tests will be carried out using consolidated solutions: total station and laser scanning.

I started working on this topic 2 years ago but several new sensors are now available on the commercial market. This has opened new opportunities for a wider user community.