Probabilistic object classifications

Researcher:
Prof. Vadim Indelman | Aerospace Engineering

Categories:

Automation, Mobility and Aerospace

The Technology

Object detection and classification is a component of situational awareness important to many autonomous systems and robotic tasks. The mobility of robotic systems is widely exploited to overcome limitations of static, one-point-of-view approaches to measurement classification. These limitations may include problems such as occlusions, class aliasing (due to classifier imperfections or objects that appear similar from certain viewpoints), imaging problems, and false detections.

The technology suggests a method for robust visual classification of an object of interest, taking into account multiple observations and viewpoints, multiple sequential images and 3D reconstruction of the scene estimated by of the camera pose.

Advantages

  • Taking into account multiple observations, sequential images and 3D reconstruction reduced the uncertainty and can confirm an improvement in robustness over state-of-the-art

Applications and Opportunities

  • Useful in any application that requires an accurate object or hypothesis classification: Augmented reality, autonomous navigation, 3D reconstruction, aerial navigation in GPS-deprived environments, satellite proximity operations, and more.
arrow Business Development Contacts
Ilia Baskin
Director of Business Development, Engineering