Theses

Signal fusion from SPID and TOF cameras

This thesis will be carried out in collaboration with the research group of prof. Franco Zappa in Micro Photon Devices.

The aim of the thesis is the data fusion of depth maps (images with distance information of every point from the camera) obtained by Time of Flight (TOF) cameras or structured light cameras (like Kinect device) with Single Photon Imaging Detectors (SPID).

This will allow a significant improvement in range data acquisitions for real-time applications in many application fields: from automotive to industrial robots, from surveillance to domotics.

Automazione di industria

L'azienda Pariani srl di San Macario di Samarate (VA) offre una tesi specialistica in un processo di Produzione e verniciatura manuale.

Prodotti – pannelli retroilluminati / applicazione campo aeronautico;

Parametri di lavorazione / misurazione conformità prodotto: coordinate cromatiche (misurata in u’, v’) e luminosità (misurata in fL)

Obiettivo dell’azienda – sviluppo processo di verniciatura automatica mediante utilizzo di attrezzature industriali e di laboratorio già presenti sul mercato, da acquistare.

 Obiettivo di progetto -  sviluppo di un algoritmo per la creazione di un filtro SW di correzione per la stampa. Il processo deve essere automatico e basato su un algoritmo di correzione immagine real-time, che agisce direttamente sugli impianti di produzione. Sarà necessario studiare il sistema ottico  e ricavare il modello matematico del materiale per regolare i parametri di processo e raggiungere il requisito predefinito.

 Risultato atteso: entro max 3 passaggi di verniciatura, direttamente in produzione ed attraverso una misurazione dei risultati di stampa in tempo “0”, ottenere dall’algoritmo di correzione dei parametri di stampa, il risultato atteso per coordinate cromatiche e luminosità pannello retroilluminato.  

Per informazioni contattare il docente o il dott. Simone Spanò:

Phone: +39 (0)331-236014

E-mail: simone.spano (at) pariani.com

Web site: www.pariani.com

 

Speckle noise reduction from laser scanner acquitions

The thesis is in the field of Computer Vision for Industrial applications.

The aim is the reduction of speckle noise from laser scanner acquisitions on steel rods at Tenaris Dalmine Company.

The system is based on a laser line projected on the object and a 3D reconstruction system based on two or more cameras. The thesis will develop an accurate calibration system to accurately determine cameras distortion parameters and their relative position with respect to the laser blade. Further geometrical constraints like trifocal tensor will be adopted in order to strength the calibration accuracy and to remove artifacts due to speckle noise.

Food Contaminants detection

Aim of the thesis is to implement and use advanced Machine Learning techniques to automatically detect contaminants (foreign bodies like plastic pieces, glass, metals) inside food and pharma in their container using X-ray Hyperspectral Imaging techniques.

The thesis will be carried out at XNEXT s.r.l. in Milan (link).

On this topic Ph.D. opportunities will also be available.

Novel SLAM

Novel Visual SLAM (simultaneous localization and mapping) algorithms, are able to reconstructs, in real time, camera ego-motion and 3D sparse map of the surrounding environment. Usually the scene contains some dominant planes, like walls, floors, building facades and roadway.The detection of these regions plays an important role in car ADAS (Advanced Driver Assistant System), Augmented reality applications and Indoor navigation. The goal of this thesis is enrich actual solutions, enhancing detection and recognition capabilities of actual systems.

 

The thesis will be developed in collaboration with ST Microelectronics AST Shared Innovation - Computer Vision Research Roadmap

Subcategories