Options
2000
Conference Paper
Titel
RecceMan - a user friendly interactive target recognition
Abstract
UAV image interpretation operators have to analyse directly transmitted real time image signals from the cruising sensor platform under high time pressure. In case of doubt about a target object to identify they need rapid access to meaningful reference images in order to meet a proper decision. Conventional reference libraries are generally very comprehensive but do mostly not offer search keys and sample pictures that are optimised to the task of image based object recognition. The system RecceMan (abbreviation for Reconnaissance Manual) is designed to bridge the gap between comprehensive but complex reference databases and quick, image oriented access. On the retrieval side specifically designed features access trees meet the demands of image based recognition. Dedicated trees can be composed and selected for various reconnaissance purposes and sensor configurations. The tree structure reflects the granularity of recognizable object features according to the image Interpretability Rating Scales (IIRS). Retrieval leads to object data sheets that can get configured dynamically in order to focus the interpreters attention onto the most significant object views with respect to the spectral sensitivity and looking angle of the sensors. RecceMan will be integrated into the Ground Station of the German Army's CL-289 UAV reconnaissance system.