Bachelor Thesis
3D Object Tracking and Position Estimation with multiple USB Cameras
Various mobile robot scenarios require tracking of objects - such as the robot itself - in 3D space. Today's commercially available tracking systems are designed to be robust against distracters, changing environments, or lightning conditions, tolerate object deformations, and ideally track multiple unmarked objects at high frame rates and high precision. These features come for a price: such systems cost serious amounts of money. Many tasks, however, do not need such high-quality features. In simple scenarios it is often sufficient to label objects with one or multiple active markers, such as blinking LEDs, which simplify tracking significantly. Furthermore, most experiments in academia get performed in well-known static environments. For such tracking situations, standard Computers equipped with USB cameras are sufficient, which reduces the cost for such a system drastically.In this Bachelor-Thesis the student will use vision information from two or more standard USB webcams, develop a "marker" for tracking (e.g. a blinking LED), and develop software to track this marker in 3D (position only) or 6D (position and orientation). The student shall design a simple calibration method that allows moving the tracking system to new locations with minimal effort for recalibration.This system will be used to track small hovering helicopters that ultimately shall learn to self-stabilize their flight based on on-board sensors.
You will need solid programming experience (preferably in C) to solve this assignment.
Student:
Stefan Maier
Advisor:
