Object Detection and Fusion in a Camera Network
This course is a lab project at the Faculty of Computer Science. Students from other faculties are welcome.
Intended Participants | Master Students |
Instructors | Marko Ristic |
SWS | 2 |
Credits | 6 |
Languages | English |
Required Knowledge |
|
Desired Knowledge |
|
Project Description
General description
Distributed state estimation and localization methods have become increasingly prevalent in modern networked tracking systems. For example, a vehicle may be tracked by multiple external sensors that different parties operate. To combine these sensor measurements, their data is typically gathered centrally and fused to produce a location estimate more accurate than one obtainable from only a single sensor. This project focuses on the implementation of these distributed localization schemes.
The project involves calibrating and using a small-scale camera network of Raspberry Pi nodes to track a moving object. Each node should process its camera data to detect the object and a distributed state estimation method should be used to compute an estimate of its location using the measurements from all the nodes. Different estimation methods exist and use different models that capture how an object moves; it will be up to the students to choose which of these methods and models they implement and how to evaluate them.
Project goals
- Calibration and setup of a camera network
- Image processing for the detection of a moving object
- Implementation of one or more distributed localization methods
- Evaluation of implemented methods
Subtasks
- Design and implementation of the hardware communication network
- Implementation and justifying of algorithms for object tracking and sensor fusion
- Choosing and justifying evaluation techniques of performance and accuracy
- Documentation of project
Registration
Via email
Please provide the following information
- Motivation for participating in the project
- Your studies and background
You can also apply as a team.
We will select a group of students on 04.04.2022.