Have your say
Video Surveillance: new challenges
By Lucian Ciobanu*
I have been working at INESC Porto for over eight years now and I accepted this invitation to write something about myself, about my work at the Multimedia and Telecommunications Unit (UTM) and present my view on the research I have been involved in over the years.
The relaxed atmosphere, spirit of mutual aid and the exchange of scientific knowledge between colleagues at INESC Porto are all positive factors for me, but it was the projects I have been involved in that have really contributed to my scientific development. I have been involved in various areas of multimedia and the research has been diverse, ranging from video indexing systems and small multimedia databases to distribution systems, dynamic adaptation and protection of audio-visual contents. I later began a PhD in distributed video coding looking at video surveillance with high density video cameras and focusing on the global efficiency of coding, given the visual redundancy between various views.
I recently began working on a National Strategic Reference Framework (NSRF) project, ROBVIGIL. The project aims to develop an innovative surveillance robot capable of performing tasks that are potentially dangerous for humans and it is therefore an important addition to security systems. The project uses multiple cameras in video surveillance with special emphasis on camera mobility and autonomy. Although the project is still within a similar area of research, it presents a new challenge for me, moving from video coding to tracking and identifying people with video equipment.
The video processing that I am involved in, with my colleague Pedro Carvalho, has motivated me and certainly made me more scientifically curious. The project aims to create a robot that is able to perform tasks under specific conditions such as in total absence of light and therefore, for the first time, it requires more sophisticated equipment. A thermographic camera will be used to capture infrared radiation from the heat emitted from the human body. An infrared light projector (which is invisible to the human eye) will also be used to discreetly light up the surveillance area because the images captured will be processed in real time using a high definition dedicated camera.
The challenge was acquiring technical and scientific knowledge about the “world of infrared”. This was necessary in order to adapt image processing techniques that had previously been used in conventional images. I think it is an interesting experiment that will provide new perspectives in this area.
In addition to partners from the private sector including Clever House – Sistemas Inteligentes Lda, Strong Segurança S.A. and Sinepower Consultoria Projetos de Engenharia Eletrónica Lda, I have had the pleasure of working with colleagues in the area of networks from UTM and with researchers from the Robotics and Intelligent Systems Unit (ROBIS). I believe this inter-unit collaboration is really beneficial and we have exchanged productive ideas in relation to the robot’s specifications.
*Collaborator at the Telecommunications and Multimedia Unit (UTM)