An FPGA-based motion-vision integrated system for real-time machine vision applications

Intelligent industrial robots must respond rapidly to environmental stimuli, and therefore the integration and real-time processing of motion-vision functions are of utmost importance. Robot decision making and motion control are commonly based on visual information in industrial machine vision applications. However, conventional processors are inadequate for such real-time processing since they are not designed primarily to perform repetitive operations on a large amount of data. In this paper, we propose a new embedded system that performs motion-vision functions in real-time using a dedicated, integrated processor. By employing both conventional and a dedicated processor, we were able to design a motion-vision integrated system that is both efficient and versatile.