Context and Background:  Tracking systems are essential in computer vision, with applications in surveillance, health-care, etc. Recent developments in video acquisition hardware and the consequent relaxation of its price have made possible a deployment of multiple cameras. The availability of multiple points of view and the redundancy of information allow a more precise monitoring of complex scenes.

Problem Statement: Object/Human tracking using images is one of the necessary research topics in dynamic scene analysis, since tracking is typically the first step before applying higher level scene analysis algorithms. This proposal is concerned to develop a multi-camera tracking algorithm. The hardware of the system includes two spherical robots, each of them equipped with several cameras distributed over their surfaces. Detection and tracking object/human in this system is challenging, due to incorporation of several cameras.

Outcomes and Deliverables:

The major outcomes of the project are:

• Multi-view based detection of moving targets (human/object)
• Tracking of detected targets
• Programming skills using Matlab and C++

The deliverables of the projects include: 

• Conduct a detailed survey of moving target detection and tracking using multiple views.
• Development of detection and tracking techniques of moving targets using captured data from the multiple cameras mounted on the two spherical robots.
• Simulation and experimentation on hardware to be available at the ISR.
• Dissemination and Technical report on findings.

Tasks:

1. Calibration of the cameras
2. Registration of the image data
3. Design and develop target feature extraction
4. Design and develop Bayesian tracking method for human/object tracking
5. Simulate and demonstrate a working prototype
6. Experimentally demonstrate simulation results with comparisons against ground truth
7. Technical report and Dissemination

Pre-requisites Required for Student:

• Programming Skills
• Interest in Computer Vision and Image Processing

Supervisors:

• Professor Jorge Dias; Institute of Systems and Robotics, University of Coimbra, Jorge@isr.uc.pt
• Hadi AliAkbarpour; Institute of Systems and Robotics, University of Coimbra, Hadi@isr.uc.pt