Object detection using event camera for control and estimate of the trajectory of a UAV
Loading...
Date
Authors
Μαχαλιώτης, Ιωάννης
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Interesting technological inventions are being produced every year around the world,
in order to solve difficult every day problems and make life easier, explore the world, or
other planets and all of them are based on software and hardware engineering. Access to
this type of powerful tools is sometimes limited, due to lack of sufficient computing power
to support all of their abilities. It wouldn’t be an exaggeration to say that one of the
best inventions of the 21st century is the event camera. This particular tool has helped
a lot of scientists and researchers to accomplish and fulfill a lot of goals on the basis
of Automation Control needs, and specifically on creating real-time systems that can
interact with the surroundings and give only useful feedback to humans, to understand
more about the world.
On this thesis, serious progress has been made on using the event camera to analyse
the dynamic environment and extract useful features, which can be used for Unmanned
Aerial Vehicles (UAV) navigating, trajectory estimate and control in the future. Dynamic
Object Detection is one of the most important and also challenging work on the field
right now. Also, responsible attempts of reducing input data alongside maintaining low
computational cost have been made, achieving high score of dynamic object detection,
which is the main idea. Lastly, for event processing a new algorithm is presented which
is inspired by the popular theory of relativity and the Hungarian method, even though
no knowledge of the last has been used for it.
Description
Keywords
Dynamic object detection, Event camera