1: Has the system already been tested in an industrial setting?
Yes, the system has been tested for more than three years at the FORD Factory in Valencia, Spain. It is installed within the production line and completely inspects as many as 1,700 cars a day.
2: What kind of technology is being used in the inspection tunnel?
The system is based on artificial vision, which processes images obtained from CCD area cameras. In addition, car bodies are illuminated using structured light, created by the movement of high frequency fluorescent tubes.
3: Is high level training required in order to use the inspection tunnel?
The tunnel has been developed as a completely automated system, with no required intervention by workers. Also, the system integrates self-tuning and learning processes that cope with environmental changes in industrial conditions. In fact, maintenance is required only for the exchange of exhausted tubes and broken cameras, if necessary.
4: Why are learning processes required?
The system is very precise, detecting micro-defects over the whole car body. However, the inspection tunnel, as well as other large mechanical systems, has tolerances usually higher than the small size of the defects. In addition, the positioning of cameras, tubes and, in special car bodies, introduces uncertainties that have to be evaluated and compensated for. Learning systems have been designed in order to overcome such situations.
5. Is it possible to recover information about a car body that has been inspected in the past? For how long is such information available?
The system has been designed in order to store important images and information about the previous 10,000 car bodies. However, a back-up procedure has also been introduced in order to store historical information.