Automated object quantity estimation for Radar Imagery

Vadum will develop a deep learning-based computer vision technique to automatically and accurately estimate the number of objects in a synthetic aperture radar (SAR) image. The technique learns on key features of SAR and optical images while being resilient to a wide range of SAR image resolutions, collection geometries and object poses. The analysis architecture is extensible to images collected from other radar imagery sensors, as well as electro-optical (EO), Infra-red (IR) and Hyper-Spectral Imaging (HSI) sensors of varying fidelity. The technique will estimate the number of objects in a SAR image more rapidly than a human analyst. The automation of this process will enable an analyst to focus on more complex operations such as situational understanding of a scene and recommending timely modifications to orders-of-battle.