Latest question:
How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!



Auto Assembly
The perception of quality is often very subjective. While consumers today are certainly more sophisticated than in the past, especially with research being more easily attainable via the Internet, a deal often will hinge on the sound of a slamming door.
EDAG (Fulda, Germany), a worldwide developer of production concepts, primarily in the automobile and aerospace industries, teamed with Allied Vision Technologies (Stadtroda, Germany) to develop an optimal fit program. Normally, attached components are assembled using gauges for measurements. The problem is that the gauges assume that the components are in a constant, rigid position for installation. That often requires substantial manual adjustments because uniform distribution of gap dimensions is difficult to attain without extensive—and expensive—labor.
To solve the problem, EDAG worked with Allied Vision to develop an intelligent sensor to use with its Best-Fit image-processing system. In door installation, for example, the sensors determine the dimensions of the gap and transition for each manufactured door along with the cut-out in the side wall of each vehicle. It then calculates the optimal position of installation of that door with the corresponding car body. Finally, robots target the position and install the door.
The system requires six to 11 EDAG VarioGauge V4 sensors, depending on the design of the attached part, installed on the gripper of the robot. In addition to electronic control and monitoring systems, the system requires a laser source and an Allied Vision camera—in this case from the Marlin family.
For better integration, the housing of the camera was removed and the board-level version was selected. With a resolution of 1.4 megapixels, the camera permit precise gap measurements. They are equipped with a FireWire IEEE 1394a interface, which can be used to activate all the cameras in the Best-Fit system via a broadcast trigger. With the camera in the deferred mode, the system can call up images of each camera for analysis.