This document describes the design and development of COBRAKIN, a device conceived within the COBRA Research Project, whose purpose is acquiring in real time a full 3D vision in 360°. This feature can be exploited for the monitoring and the surveillance of indoor environments (in our case museum spaces) in an innovative way compared to the actual security devices. COBRAKIN is designed with 4 Microsoft Kinect set at 90 degree one with the other, and connected to two mini-PC (2 Kinects connected to a PC-SERVER and 2 Kinects to a PC-CLIENT) that communicate through a network connection and a network protocol developed for this aim. The infrared devices inside the Kinects, coupled to the HD cameras, allow to scan in real time the surrounding environment in three dimensions, thanks to a process of depth measure called “structured light”. The software used for the development of the user interface application and for the computer vision algorithms is based on the Processing programming language, and on further open-source tools and libraries. Several algorithms have been implemented to introduce useful features for the video surveillance. For example, the Depth Threshold to set a security distance against a target, the Frame Compare to detect only the subjects in movement inside the scene, and the Background Subtraction to detect changes compared to the initial scene.
Questo documento descrive la progettazione e la realizzazione di COBRAKIN, un sensore sviluppato all’interno del Progetto di Ricerca COBRA, che ha lo scopo di acquisire in tempo reale una visione 3D a 360°. Questa caratteristica può essere sfruttata per il monitoraggio e la sorveglianza di ambienti interni (nel nostro caso spazi museali) in maniera innovativa rispetto agli attuali dispositivi di sicurezza. COBRAKIN è formato da 4 Microsoft Kinect disposte ad angolo retto l’una con l’altra e collegate a due mini-PC (2 Kinect ad un PC-SERVER e 2 Kinect ad un PC-CLIENT) che comunicano tra loro per mezzo di una connessione di rete e di un protocollo sviluppato ad hoc. I sensori infrarossi contenuti nelle Kinect, abbinati alle telecamere HD, permettono di scansionare in tempo reale l’ambiente circostante in tre dimensioni grazie al processo di misurazione della profondità detto “a luce strutturata”. Il software utilizzato per lo sviluppo dell’interfaccia utente e degli algoritmi di computer vision si basa sul linguaggio di programmazione Processing e su ulteriori tool e librerie open-source. Sono stati implementati diversi algoritmi che introducono caratteristiche utili ad un dispositivo per la video-sorveglianza. Tra questi il Depth Threshold per impostare una distanza di sicurezza rispetto ad un target, il Frame Compare per rilevare esclusivamente figure in movimento nella scena, e il Background Subtraction per determinare cambiamenti rispetto alla scena iniziale.
COBRAKIN: sviluppo di un sensore per il monitoraggio all'interno dei musei
Nuvoli, Marcello;Di Frischia, S.;Guarneri, Massimiliano
2017-10-01
Abstract
This document describes the design and development of COBRAKIN, a device conceived within the COBRA Research Project, whose purpose is acquiring in real time a full 3D vision in 360°. This feature can be exploited for the monitoring and the surveillance of indoor environments (in our case museum spaces) in an innovative way compared to the actual security devices. COBRAKIN is designed with 4 Microsoft Kinect set at 90 degree one with the other, and connected to two mini-PC (2 Kinects connected to a PC-SERVER and 2 Kinects to a PC-CLIENT) that communicate through a network connection and a network protocol developed for this aim. The infrared devices inside the Kinects, coupled to the HD cameras, allow to scan in real time the surrounding environment in three dimensions, thanks to a process of depth measure called “structured light”. The software used for the development of the user interface application and for the computer vision algorithms is based on the Processing programming language, and on further open-source tools and libraries. Several algorithms have been implemented to introduce useful features for the video surveillance. For example, the Depth Threshold to set a security distance against a target, the Frame Compare to detect only the subjects in movement inside the scene, and the Background Subtraction to detect changes compared to the initial scene.File | Dimensione | Formato | |
---|---|---|---|
RT-2017-31-ENEA.pdf
accesso aperto
Licenza:
Creative commons
Dimensione
2.36 MB
Formato
Adobe PDF
|
2.36 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.