An Open Framework for Computer Vision Systems, with Application to Intelligent Load Handling

The proposed effort attempts to drive down the cost of Computer Vision systems by advancing a component-based software framework that will allow system integrators to customize vision-based applications from prefabricated, interchangeable parts. This component-based approach provides end-user flexibility, while affording large economies through software reuse. The innovation is the development of an Open Graphical Framework that provides the benefits of graphical programming without tying the user to a proprietary solution; i.e., vision components developed within the open framework can still execute outside of the environment. This attractive feature is the key differentiator that sets the proposal apart from other graphical programming packages, such as LabView or Simulink. Proof of concept shall be evidenced by the development of modular components that form the basis of a vision-based pallet detection system. Specifically, this system will be able to identify and locate munitions pallets for automatic retrieval and/or placement by the Smart Crane Ammunition Transfer System (SCATS) technology demonstration platform. Software reuse will be demonstrated by re-configuring the vision system for different hardware platforms and for different missions.