OMRON Creates a Camera Network for Its Cobots
OMRON developed new vision and image processing capabilities for their collaborative robots (cobots), benefiting the machine tending and quality assurance divisions in manufacturing facilities.
Since last year, OMRON has been developing and fine-tuning a way for their cobots to see. Their solution consists of an onboard camera, networking capabilities, and imaging software.
The onboard 5-megapixel camera acts as the “eyes” for the collaborative robot (cobot). It is suitable for reading barcodes, QR codes, DataMatrix codes, color identification, and other measurement and inspection tasks. Adding the camera to the cobot directly reduces startup times, as the camera suffers fewer calibration issues.
Another advantage to the OMRON system is the ability to network cameras. While one camera is already mounted on the cobot, other cameras can be mounted around the workspace. These cameras serve several purposes.
First, they can perform parallel inspection tasks, where the cobot checks some measurements and a stationary camera elsewhere checks others, speeding up the entire process. Additional cameras can subtract out any offsets due to misalignment through a triangulation process.
Multiple cameras that see the workpiece and each other can help eliminate error.
A cobot works together with a networked camera. Image courtesy of OMRON
With the new hardware capabilities, OMRON has partnered with Canon to offer a matching software package. The Vision Edition-T software is designed to handle the data from multiple networked cameras, perform calibrations, measurements, and inspections. The software is designed for ease of use, integrating with many other plug-and-play hardware options.
Machine tending is one of the fastest-growing applications for cobots. In the past, a human would load and unload processing equipment, often spending time waiting for the process to complete, or performing other tasks, only to rush back and remove parts from the process equipment.
A cobot can hurry up and wait. Cobots, such as those designed by OMRON, can load, unload, arrange, and orient parts between different processing steps. Then, it can wait until it is needed again.
An OMRON robot performing machine tending tasks. Image used courtesy of OMRON
For example, suppose small copper plumbing fittings have been dropped onto a conveyor after leaving a deburring operation before their final inspection and packaging. The packaging step requires that all of the odd-shaped pieces be oriented in a specific way, and there seems to be no way to do this mechanically.
In the past, workers would have to manually position each part to ensure proper packaging. Instead, a cobot, such as this OMRON system, can identify the optimal way to move each fitting.
Quality assurance and inspection tasks are also a good fit for this cobot system. With multiple cameras, personnel can perform inspections quickly and accurately. The cobot can manipulate parts and perform the finer inspection, while the network cameras handle bigger inspection tasks.
When I was a child, I toured a large potato chip factory and passed through an area where pretzels were inspected. Workers, clad in hairnets, stood on both sides of a conveyor with large hooks to remove the deformed, burned, or undercooked pretzels. Back then, I knew nothing of quality assurance, machine vision, or cobots, but one thing was for sure—I didn’t want to have to stand and watch pretzels all day.
With the OMRON system, the cobot could be configured to identify the defective pretzels and remove them without requiring the hairnetted people and hooks.
OMRON’s new “eyes” have the potential to improve machine tending and inspection tasks. The high-resolution camera mounted on the cobot reduces calibration and integration issues. Meanwhile, the network camera capabilities enhance the accuracy, and the Vision Edison-T software performs the measurements and manages the system.