The vision hardware includes a TM-200 video camera from JAI Pulnix Inc. (Sunnyvale, CA); an RL36120 red LED ring light, high-flex camera power cable, high-flex camera coaxial cable and power supply from Advanced Illumination Inc. (Rochester, VT); and a 4sightII computer from Matrox (Quebec, Canada).

The Matrox computer includes an Orion PC104+ frame grabber, which is used to capture visual input. Other major workcell components include a custom end-of-arm-tool, pneumatics package, robot riser, two-part rack locators, divider sheet storage rack, interface panel and a safety fence.

The robot is programmed using FANUC Robotics' Windows-based vision program, with the dedicated computer providing easy access to the vision system, both for troubleshooting and programming new parts geometries.

Parts are located by using FANUC Robotics' geometric pattern matching algorithm, with the vision system connected directly to the robot via Ethernet through the robot server. The key parameters used in the application are size and contrast. These characteristics are combined to create an overall "score," which is used to determine if a part has been found or not.

For example, during testing the vision system identified "phantom" parts that were caused by oil rings the gears had left on the plastic board. The problem was solved using the contrast locator parameter. On a contrast scale of 1 to 100, the oil ring scored a contrast value of 44, versus a contrast value of 95 or more for actual parts. Phantom part finds were therefore easily eliminated by adjusting the vision recognition system.

Initially, there were also problems with the vision system locating gears that were one level lower than that which the robot was unloading. Specifically, the edges of lower-level parts located on the periphery of the pallet were visible from below the plastic divider. The problem was solved using the size locator parameter to ignore parts that were not at least 95 percent of the trained part size.

After the system was installed, Len Industries also had FANUC Robotics add part orientation detection so the robot will recognize any gears that have been placed in the bin upside down. This is vitally important because any gears entering the machining center in the wrong orientation will damage the tool. Using a clockwise pointing part feature, the vision program is able to identify upside down parts and load them into a reject chute, where they are then manually loaded to the tool.

FANUC also upgraded the vision display to show what edges the vision system is using to locate a part. Specifically, the company added red and green indicators on the image in the vision interface. The green light shows a positive edge find, while red edges indicate where an edge was expected, but not found.

Ultimately, the workcell met every one of its operational goals. After 4 months of operation, the system increased throughput by 5 percent and decreased labor costs by 30 percent, with a system uptime above 90 percent.

For more on robotic parts handling, visit www.fanucrobotics.com, call 800-47-ROBOT or eInquiry 1.

For more on cameras, call 800-445-5444, visit www.pulnix.com or eInquiry 2.

For more on illumination hardware, visit www.advancedillumination.com, call 802-767-3830 or eInquiry 3.

For more on robotic controllers, call 514-822-6000, visit www.matrox.com or eInquiry 4.