During the manufacturing of printed circuits, the process of track drawing is performed. This process consists of attacking a plaque, which is completely covered by copper, with an acid that dissolves all the copper in between the painted tracks. The problem is that sometimes some tracks are also deleted, or broken by the acid, and some holes or other kinds of defects can take place. We successfully designed and implemented a system that uses Computer Vision techniques for the automatic detection of the defects present in the plaque. Our system uses the original design of the plaque as a template, and it characterizes the defects into short-circuits, holes or broken tracks with a high accuracy.