VideoSurv - video surveillance

The project VideoSurv is situated in the context of video surveillance, particularly in the domain of the security of goods and services. The augmented utilization of video surveillance cameras generates immense quantities of data. The goal of the project VideoSurv is to automate the extraction of video sequences capturing relevant events. The kea idea is to fuse the stream of images coming from different types of cameras (thermic cameras, cameras measuring the distance) in order to facilitate the automatic annotation of the video streams and therefore enable the operators to easily find important sequences.

Keywords

video surveillance, data annotation

Outcomes

In order to validate our approach, two demonstrators, which adopt the same approach, will be developed. They will show the relevance of the combined utilization of at least two types of cameras in the framework of video surveillance. Demonstrator HE-FR uses a color camera and a 3D camera (Kinect), which monitor a waiting room and signalize the entry/exit of people, as well as keep the count of the number of visitors. Demonstrator He-Arc uses a thermic camera and a color camera in order to monitor the displacements of cars and people in a parking lot.

Website of the project

 

Project Information

This project is elaborated in collaboration with He-Arc (Haute Ecole Arc - College Arc).