RIT Team Is Developing Computer Vision Technology to Improve Aerial Tracking

Facebook X LinkedIn Email
ROCHESTER, N.Y., Jan. 9, 2019 — Researchers at Rochester Institute of Technology (RIT) are developing a visual tracking system to more accurately locate and follow moving objects under surveillance. Using deep learning, the system could generate more reliable readings of moving objects, even if objects become obscured or change patterns or direction.

Professor Andreas Savakis is heading the three-year, $250,000 project funded by the Department of the Air Force’s Materials Command/Systems and Technology. The RIT system is one of several interrelated projects that will use deep learning to refine the computing algorithms and visual tools needed to follow the movements and understand the activities of objects under surveillance. Savakis’s team will produce a prototype tracking system, building upon earlier research conducted on video analytics.

Although many systems have the ability to recognize and classify various objects, it is essential to track the objects under variations in illumination or appearance to re-acquire them when they are occluded by other objects, Savakis said. Some challenges to current imaging technology and the need for improvements arise from objects seeming to change size and appearance because of distance and perspective, or lighting that can produce different colorings of objects. The researchers will be using deep neural networks and visual tracking algorithms to train their system to understand these distinctions. 

The research has potential applications in autonomous navigation, drones, traffic monitoring, safety and security, disaster response, and human-computer interaction.

Published: January 2019
deep learning
Deep learning is a subset of machine learning that involves the use of artificial neural networks to model and solve complex problems. The term "deep" in deep learning refers to the use of deep neural networks, which are neural networks with multiple layers (deep architectures). These networks, often called deep neural networks or deep neural architectures, have the ability to automatically learn hierarchical representations of data. Key concepts and components of deep learning include: ...
surveillanceResearch & TechnologyeducationRochester Institute of TechnologyRITImagingAmericasfundingLight SourcesSensors & Detectorscomputer visiondeep learningdeep neural networksdefensesecuritymilitary

We use cookies to improve user experience and analyze our website traffic as stated in our Privacy Policy. By using this website, you agree to the use of cookies unless you have disabled them.