But Braintech Inc. believes it has a better idea. Babak Habibi, president at the North Vancouver, British Columbia, Canada-based company, says Braintech’s SC3D vision technology can perform 3D part location using a single image from just one camera.
SC3D—which stands for Single Camera 3 Dimensional—works by focusing on key features, or “landmarks” on a target part, and applying a method called “perspective distortion,” says Habibi. And under terms of a partnership with robotics vendor ABB Inc, SC3D technology has been applied for what Habibi calls “semi-structured” or “semi-constrained” robotic bin-picking applications at numerous North American automotive factories.
According to Jerry Osborn, vice president and general manager of ABB’s Robot Automation Division, in Auburn Hills, Mich., around 60 out of 100 Braintech vision-guided ABB robot applications in the field qualify as “semi-structured bin-picking.” The customer list includes DaimlerChrysler, Delphi, Ford, General Motors, Honda, Nissan and Toyota.
In these applications, thanks to SC3D robot guidance, parts can be loosely constrained in bins for robot picking, often eliminating the need for expensive precision dunnage, says Osborn. In major automotive programs, the savings can amount to “hundreds of thousands of dollars, just for one set of dunnage for one set of parts,” he contends. Another major savings comes through labor reductions, Osborn adds, “because we’ve been able to break into applications that were done manually that weren’t feasible robotically before.”
One end-user that is pleased with the technology is Bob Welch, assistant manager of engineering at Toyota Motor Manufacturing West Virginia (TMMWV), a power-train plant in Buffalo, W. Va. The Toyota plant installed five SC3D-guided ABB robots last year as raw material unloading systems for cylinder blocks and heads. The systems have helped boost productivity at TMMWV by reducing labor costs in some machining areas, says Welch.
“The beauty of the ABB/Braintech solution is that it’s 3D vision with one camera,” Welch observes, “so the complexity to the user is minimal, and you don’t have a two-camera system that you’re trying to keep calibrated.”
See the story that goes with this sidebar: In Search of the Holy Grail