One of the main application area of computer vision lies in the industrial area. Automation of processes is done be replacing the human operator or inspector by systems that use computer vision to solve the same task cheaper and often even better. The motivation for installing such systems lies in the economy and the quality assurance by using this systems. But there are also other problems that should be solved even if there is no economic benefit. Therefore it is of interest if methods, that were intended to work in "practical" applications could also be used in other areas.

This talk shows methods that were developed to solve the automatic calibration of watermeters and how their development influenced two non-profit applications: the classification of portrait drawings of the 18th century and the classification of archaeological sherds. These two selected applications in the field of art history and archaeology show, that "traditional" sciences do have a need for state of the art computer vision techniques and that the results obtained are of interest in this research communities.