Integration of COTS UAS With Multispectral Imaging Sensors to Detect Camouflaged Targets and Battlefield Anomalies

In response to the extended use of tactical and theater-level UAS for reconnaissance and surveillance in the modern battlefield, ground forces are increasing their efforts to hide their assets, using camouflage or exploiting terrain and vegetation. In the last decade, the technology of multispectral imaging has evolved, providing compact and low-cost sensors that can enhance the capability of tactical UAS for defeating camouflage and detecting battlefield anomalies that are not visible with EO/IR sensors. In contrast, multispectral sensors provide imaging in multiple narrow bands in the visible and IR spectrum. Computing a difference in absorption and reflectance of different materials in these narrow bands and then fusing these multispectral data in a certain way creates an opportunity to detect even the camouflaged targets. This thesis compares performance of a COTS multispectral imaging sensor integrated with a tactical UAS to detect camouflaged targets and battlefield anomalies against EO/IR sensors. In doing so, it introduces a new image data fusion methodology, examines its effectiveness compared to several other methods, and evaluates its overall performance. The promising results achieved in processing imagery data collected during several field testing campaigns employing an eVTOL UAS equipped with a MicaSence five-band sensor suggest a necessity to explore this concept further, to include a larger database encapsulating different contemporary camouflage techniques.