Translator Disclaimer
12 October 2010 Application of scene understanding to representative military imagery
Author Affiliations +
Scene understanding (SU) is a high priority in many areas. Currently many SU algorithms are developed using imagery which is often captured in constant, well lit environments with low clutter and not affected by noise, compression or bandwidth artefacts. The initial research addressed how SU can assist automatic identification, semantic tagging and tracking of an object in a scene. However, it became apparent that current algorithms and software are unable to successfully process typical military imagery. Consequently, research was undertaken to assess how well current SU algorithms process imagery captured by a variety of typical military imagers. The imagery was chosen such that it covered a variety of scenarios and applied to a range of algorithms. It became apparent that the many algorithms experienced difficulties in processing the typical military imagery or had other drawbacks such as computational cost, which impacts on military utility. The National Imagery Interpretability Rating Scale (NIIRS) was used in order to help explain the general quality of military imagery and the analysis tasks which are expected to be carried out on it.
© (2010) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Natalie Dyer "Application of scene understanding to representative military imagery", Proc. SPIE 7838, Optics and Photonics for Counterterrorism and Crime Fighting VI and Optical Materials in Defence Systems Technology VII, 78380L (12 October 2010);

Back to Top