Real-time Single Chip Target Acquisition and Tracking for sUAV Gimbal

Over the last several years there have been significant advances in algorithms involving targeting tasks onboard airborne gimbals. These tasks include moving target indication (MTI), discriminative correlation filters (DCF) for click-to-track correlation tracking, as well as target classification via neural networks. While all these algorithms provide significant functionality, they have traditionally come at the cost of substantial processing hardware requirements. With increasing interest in related algorithms being made available to mobile devices, there are now computing solutions that can realize these advanced capabilities while simultaneously deploying them in low SWaP-C footprints compatible with classes of gimbals deployed to small unmanned air vehicles. In this effort Toyon is targeting a production tracking solution towards a 5â€Â gimbal currently in development. Additionally, within the same ARM processor family used in the 5â€Â version, Toyon will begin development of a tracking solution for a 3â€Â gimbal. These solutions will share the same algorithm and similar software code base with a variation in the set of capabilities. Namely, the 5â€Â will focus on search tasks with MTI and then also host a correlation tracking function. The 3â€Â will be more focused on click-to-track, but contain related elements, such as stabilization.