Intelligent camera cloud computing

In many industrial automation processes, quality monitoring is based on visual inspection. An example is flat-steel manufacturing, where surface defects need to be monitored. Currently this is performed by extensive image logging and off-line human interpretation. More intensive automation requires gathering quantitative information, but this will easily require a very expensive camera, if feasible at all. Furthermore it poses a burden in guaranteeing proper lighting conditions. The paper describes experiments with a cloud of cameras. It is shown that the use of multiple cameras will not only raise the robustness of the system but also the precision of the measurement. This allows measuring micrometer defects from a distance using a cloud of cheap vision sensors.