Publication Date:
2012-02-28
Description:
Finding explosive threats in complex environments is a challenge. Benign objects (e.g. rocks, plants and rubbish), ground surface variation, heterogeneous soil properties and even shadows can create anomalies in remotely sensed imagery, often triggering false alarms. The overarching goal is to dissect these complex sensor images to extract clues for reducing false alarms and improve threat detection. Of particular interest is the effect of soil properties, particularly hydrogeological properties, on physical temperatures at the ground surface and the signatures they produce in infrared imagery. Hydrogeological variability must be considered at the scale of the sensor's image pixels, which may be only a few centimetres. To facilitate a deeper understanding of the components of the energy distribution, a computational testbed was developed to produce realistic, process-correct, synthetic imagery from remote sensors operating in the visible and infrared portions of the electromagnetic spectrum. This tool is being used to explore near-surface process interaction at a fine scale to isolate and quantify the phenomena behind the detection physics. The computational tools have confirmed the importance of hydrogeology in the exploitation of sensor imagery for threat detection. However, before this tool's potential becomes a reality, several technical and organizational problems must be overcome.
Permalink