imade
Imagery-based Missile Attitude Detection and Estimation

IFT’s IMADE uses perfect a priori knowledge of target geometry to determine target attitude/pose. This software solution is applied to ground-based images of in-flight missiles in arbitrary cluttered backgrounds. It takes advantage of state-of-the-art segmentation algorithms, known 3D models of the target, and a single view of the target in motion. IMADE cultivates a rich dataset to enable both model-based and deep neural network-based methods. It utilizes classical and machine learning-based segmentation, and image-based & model-based target attitude determination. It is ready to be expanded to a multi-view solution.

METHOD: Shape-preserving segmentation, 3d model rendering & fitting

KEYWORDS: Attitude determination, pose estimation, image segmentation

TECHNICAL SPECS: Pytorch, Pytorch3D, Unity

RESEARCH: Flying Missile Attitude Determination from Monocular Ground Based Imagery with a Priori Knowledge, to be submitted to IEEE Aerospace Conference and/or IEEE Aerospace and Electronics Magazine

Download the IMADE Brochure

DOWNLOAD
Style Settings