Faculty of Engineering, LTH

Denna sida på svenska This page in English

Digit@LTH: Events

MSc. presentation by L. Hansson & M. Lindeberg: PTZ Camera Tampering Correction Using IMUs

Axis camera at an airport


From: 2022-05-31 14:15 to 15:00
Place: Seminar Room KC 3N27
Contact: anton [dot] cervin [at] control [dot] lth [dot] se
Save event to your calendar

Ludvig Hansson and Måns Lindeberg are defending their Master's thesis at the Dept. of Automatic Control.

Where: Seminar room KC 3N27
When: May 31, 14:15-15:00
Authors: Ludvig Hansson and Måns Lindeberg
Title: PTZ Camera Tampering Correction Using IMUs
Advisors: Anton Cervin, Dept. of Automatic Control, and Marcus Romner, Axis Communications
Examiner: Kristian Soltesz, Dept. of Atutomatic Control

The pan-tilt-zoom (PTZ) camera is a type of camera that cover large areas, up to 360°, in addition to zoom. Some of these cameras are susceptible to tampering, which means that an additional angle measurement feedback is necessary to detect such attempts. This thesis has evaluated the possibility of using MEMS-based IMUs to solve this problem. An advantage of such a system over other solutions, such as optical and magnetic encoders, is the decreased cost and complexity. The difficulty of using IMUs is that they are susceptible to drift in horizontal rotations in pan and requires advanced signal processing to be reliable in demanding environments. Results showed that it was possible to achieve tampering feedback using two IMUs, one on the fixed part of the camera and one on the moving part of the camera. It was then possible to track and correct tampering attempts with an average error of 0.0 degrees in pan and 0.7 degrees in tilt. The solution works when the camera is mounted on fixed objects and moving objects. Notably, the solution proved to be much cheaper than alternative methods, such as those using magnetic and optical encoders. Additionally, it proved to be robust against vibrations and shocks and within the performance requirements of the tested model. Additional measures were highlighted for the system to track absolute orientation. For example, to be able to move the camera to pre-set positions. However, further work is needed to increase robustness in applications where the camera is mounted on objects that can rotate in roll or pitch.