Wednesday, 21 Feb 2018
You are here:

Subscribe to our Newsletter

How many eyes has a typical person? (ex: 1)

Twitter Feed

EMail Print


Dr. Pavel Korshunov (EPFL) visiting Prof. Jean-Luc Dugelay (Eurecom)
14.04.2014 - 19.04.2014
Internal Visits


This research visit focused on defining and planning the new mini-project on privacy protection and ethical use of mini-drones in surveillance applications.
When there is a major temporary event such as concert or sport event, it is not possible to setup a fullyfledged surveillance system, hence, mini-drones can be used to survey the area and help in managing parking spaces, crowd control, etc.
For instance, to provide such surveillance in a temporary parking lot, mini-drones can be used for scanning the scene and reporting some information, including the number of free lots, identified cars that have not moved for an extended period of time, suspicious behaviour, etc. Special care needs to be taken to insure personal privacy such as protection of license plate numbers and people.
This mini-project is the result of successful continuous collaboration between EPFL and Eurecom on evaluation methodologies for assessment of visual privacy. An open call for the participation in minidrones mini-project was sent to VideoSense partners at the end of the research visit.
During the visit, Eurecom and EPFL have identified the scenario of using mini-drones for temporary car-park surveillance. To proceed with the mini-project, it is necessary to have a relevant set of video data, hence, the following sequences of an evaluation video dataset that need to be shot from minidrones are planned:
1. Scan of the parking when it’s empty
2. Scan when it’s full - without people
3. Scan when it’s full - with people
4. Scan when half full - without people
5. Scan when half full - with people
6. Scan parking with a motorcycle - without people
7. Scan parking with a motorcycle - with people
8. Mis-parked car. One car blocks entrance - bad parking. Then the drone has to go and check the Licence plate.
9. Mis-parked - bad parking.
10. Suspicious behaviour - loitering, zigzag walking
11. Suspicious behaviour - reserving the spot illegally
12. Obstacle in the parking lot
13. Fights in the parking
14. Inappropriate activities: playing football in the parking lot, skating, etc. 15. Big queue in the parking

Also, during the visit the following participatory components of this mini-project were planned for joining partners of VideoSense:
1. Generation of the video content according to the specified scenarios or new proposed scenario
2. Detailed analysis of surveillance by drones and related applications, technologies, and tools (produce a detailed report)
3. License plate detection and tracking
4. People detection and tracking (head, pose, body, face, gate, etc.)
5. Unusual events detection (loitering, zig-zag, falling, misparked cars, etc.)
6. Virtualization of the parking lot and objects in it (empty lots, cars, people, obstacles, etc.)
7. Privacy tools applied to the scenarios
8. Annotations
The mini-drones mini-project will allow partners to cooperate together and develop a set of use cases, scenarios, privacy protection technologies, evaluation methodologies, and privacy-guided and ethical recommendations for a wide-spreading drone technology. The outcome of this mini-project will contribute to the sustainability of the VideoSense project for the years after its end.
During the visit, ‘parrot’ mini-drones from Eurecom have been used for demonstration purposes. The visit has contributed to WP2, WP4, WP5, WP6 and WP7 activities.