Smart CCD Camera for Real-time Programmable Image Processing using cellular Neural Networks

In spite of the rapid advances in the speed and density of electronic components, a compact hardware package providing a complete real-time solution to vision problems remains beyond the capabilities of currently available technology. This project proposes to develop a smart CCD camera using a temporal cellular neural network architecture that can perform a variety of iconic operations in real-time. The programmable features combined with the planned execution of post processing algorithms on a conventional microprocessor will enable this smart camera module to be used as a general-purpose image processing engine capable of edge finding, motion and dept computations. As a result, the proposed system will be practical in a wide array of settings, such as advanced manufacturing, hazardous material handing, and navigation, as well in a variety of commercial, and eventually, consumer video systems. The Phase I effort will be directed towards proof of concept through system simulation of the smart camera. In addition, the architectural issues involved in interfacing with a conventional microprocessor will be investigated.