Camouflage Detection Using MWIR Hyperspectral Images

Mid Wave Infra Red (MWIR) region of the electromagnetic spectrum is a key region that is exploited for long range infrared reconnaissance and surveillance applications. Detecting enemy targets and being undetectable is the prime requirement of the modern warfare. In the present study MWIR hyperspectral imaging has been exploited for detection of camouflaged targets in the scene. Detection methodology has been developed using unsupervised target detection algorithms (Reed Xiaoli, K-Means and ISODATA). Methodology has been successfully tested in extracting camouflaged targets from the background. Performance results have been compared. The proposed methodology may in future be applied in detection and identification of gases and minerals in the field.