Video Stabilization and Target Localization Using Feature Tracking with Video from Small UAVs

<p>Unmanned Aerial Vehicles (UAVs) equipped with lightweight, inexpensive cameras have grown in popularity by enabling new uses of UAV technology. However, the video retrieved from small UAVs is often unwatchable due to high frequency jitter. Beginning with an investigation of previous stabilization work, this thesis discusses the challenges of stabilizing UAV based video. It then presents a software based computer vision framework and discusses its use to develop a real-time stabilization solution. A novel approach of estimating intended video motion is then presented. Next, the thesis proceeds to extend previous target localization work by allowing the operator to easily identify targets—rather than relying solely on color segmentation—to improve reliability and applicability in real world scenarios. The resulting approach creates a low cost and easy to use solution for aerial video display and target localization.</p>

Name thesis:degree_name

MS

Grantor dc:publisher

Brigham Young University - Provo

Author dc:creator

• Johansen, David Linn

• video stabilization
• target localization
• UAV video
• feature tracking
• mosaicking
• intended video motion
• affine model
• affine motion
• Electrical and Computer Engineering

Language dc:language

English

Repository record dc:identifier

https://scholarsarchive.byu.edu/etd/944

OAI identifier oai:identifier

oai:scholarsarchive.byu.edu:etd-1943