Human tracking robotic camera based on image processing for live streaming of conferences and seminars

Area of Science:

• Robotics and Automation
• Computer Vision
• Multimedia Systems

Background:

• Photographers face challenges in capturing footage due to space, visibility, and obstacles.
• Existing systems lack the ability to maintain focus on a subject when obstacles are present.

Purpose of the Study:

• To develop an automatically controlled robotic camera system for subject detection, tracking, and video capture.
• To overcome limitations of manual photography in difficult environments.
• To enable live streaming and recording of subjects, even with intervening obstacles.

Main Methods:

• Utilized the Viola-Jones algorithm for subject feature detection.
• Integrated Radio-Frequency-Identification (RFID) for tracking non-stationary subjects.
• Employed movable motors for camera positioning and C programming for camera protocol access.
• Implemented wireless live streaming via the XAMPP platform (X-OS, Apache, MySQL, Php, Perl).

Main Results:

• The robotic camera successfully detected and tracked subjects, ignoring intervening obstacles.
• Encouraging results were obtained in self-made scenarios, considering subject speed, distance, light, and background noise.
• Live streaming was achieved on an Android-based device.

Conclusions:

• The developed robotic camera system offers a novel solution for automated video capture in challenging conditions.
• The system enhances multimedia arrangements like live conferences and seminars by ensuring continuous subject focus.
• This technology addresses the limitations of current systems by maintaining focus despite detected obstacles.