Novel Machine Vision System Automates Onion Inspection and Sorting Process

Hyperspectral imaging, innovative soft-gripper technology, and artificial intelligence software combine to automate onion quality assurance and sorting in the Netherlands.

Labor shortage issues resulting from what’s been termed the Great Resignation have become severe. Still, companies of all types must find innovative ways to keep the lights on. When designed, specified, and integrated properly, advanced automation systems leveraging robots, machine vision, and artificial intelligence software provide an effective, novel method for automating disparate processes across myriad industries. For one of the largest onion exporters in the world, this meant finding the right system integrator and industrial automation expert.

Turning Toward Automation

The labor shortage creates issues when it comes to staffing, but many jobs have always been difficult to fill. Take the task of manually inspecting food items such as onions, avocados, tomatoes, and apples. It can be difficult to find motivated employees who can accurately identify and pick acceptable products — especially for hours on end. In addition, human employees get tired, which makes them prone to making mistakes.

As a result, companies are turning to automation at unprecedented rates. But some industries are facing other challenges, as certain applications struggle to effectively automate tasks previously done by humans. In the food industry, for example, robots don’t have the dexterity required to handle delicate or difficult-to-grasp items such as onions. To automate onion quality assurance and sorting, Wiskerke Onions, a leading onion producer located in the Netherlands, tasked industrial automation company Famateq with building a system capable of solving these problems to increase efficiency, maintain quality, and drive revenue.

Gaining a Grip

In 2019, Wiskerke Onions had Famateq install a basic pick-and-place robotic system. Recently, the company sought to automate the quality assurance and sorting processes as well.

“Ordinarily, human employees would inspect the onions for quality purposes, but the customer wanted to deploy a fully automated system that checks the onions for quality — both internal and external — while also sorting by shape or color,” said Wim de Rijder, director, Famateq. “The previous robot system struggled to handle onions at high speeds, as the gripper would often lose hold of the grip around the skin of the onion.”

In need of a soft handling, food-safe gripper that could be easily adjusted to hold different types of onions — and other types of fruits and vegetables — Famateq found the mGrip solution from Soft Robotics, Inc.  Inspired by the way in which an octopus grasps and manipulates objects, the IP69K-rated gripper uses an advanced material science–based approach to pick up products of all types safely and efficiently.

The patented elastomeric plastic grippers use food-safe materials and hollow gripper fingers with flexible holds that are pneumatically controllable at low pressure, which allows the gripper to gently grasp objects such as onions without having to change or adapt variables of the automation system.

Quality Assurance and Sorting

After large plastic bins holding approximately 1 ton (1,000 kg) of freshly harvested onions make their way into the facility, a bin tipper moves the onions onto a flat belt that transports them onto small rollers in a grid array that enters Famateq’s Master Picker machine. Spread into a single layer of various shapes and sizes, the onions pass under a hyperspectral camera that captures 30 images of each onion at 1500 fps. These images feed into a proprietary machine learning—based software that analyzes the internal and external quality of the onions based on the image set.

The standard Master Picker system is modularly expandable and has four exits that can be programmed to user needs, such as designating one or more exit for rotten onions, one for different-colored onions, one for abnormally sized onions, and so on. All acceptable onions continue on the roller conveyor and move onto the packaging machine to be shipped, but any onion deemed unacceptable is quickly picked using Soft Robotics’ mGrip gripper on a delta robot and put into one of the exits.

Innovative automation solutions like Soft Robotics’ mGrip technology are the ingredients for success at Wiskerke Onions, according to the company. Moving forward, the company says it will stay on the cutting edge of automation technology to stay ahead.

A Fast, Effective Alternative

Wiskerke Onions’ new Master Picker system can run 24 hours a day, 7 days a week and removes any mistakes that a person might make because they were tired. The system also operates considerably faster than any human can.

“Wiskerke Onions currently has one machine installed at their facility that can make 170–190 picks per minute, which far outpaces human capabilities,” said de Rijder. “As a result of the success of the initial system, the customer may look to order additional systems soon.  Beyond the success at Wiskerke Onions, de Rijder also noted that another customer in the same area has purchased a Master Picker system to automate a similar process.

“Companies we’ve worked with in the food industry have taken notice of our system and the Soft Robotics’ gripper’s ability to handle difficult and delicate food items and see this as an effective method for solving a difficult automation problem,” he said.