TOMRA Food debuts pair of new AI-assisted sorting solutions
TOMRA Food debuted two new sorting and grading solutions, each featuring increased roles for artificial intelligence, at the International Fresh Produce Association’s Global Produce & Floral Show in Anaheim, California.
The TOMRA Neon pre-grader for machine-harvested fresh blueberries and the new-generation Spectrim X series were exhibited for the first time at the show, held Oct. 19-21.
The TOMRA Neon pre-grades machine-harvested blueberries before transferring the fruit directly onto TOMRA’s KATO260 optical sorter and grader, according to a news release.
The compact, durable machine removes more than 95% of clusters and more than 90% of green and red berries, according to the release.
It has a throughput capacity of up to 500 berries per second and can maintain a speed of up to 280 berries per second.
The TOMRA Neon harnesses artificial intelligence to differentiate and remove clusters as well as undersized or unripe fruit. It was prototype-tested for two and a half years in machine-harvested conditions in North America and New Zealand, according to the release.
The new-generation Spectrim X series uses deep learning for sorting and grading precision. The Spectrim X is series is TOMRA’s first solution to utilize the company’s LUCAi Deep Learning AI platform, introduced in 2017, according to the release.
Deep learning technology uses pre-trained models to teach computers how to process data such as complex patterns in photos. Spectrim X series assesses thousands of high-resolution fruit images every second, then cross-references those images with data to make grading decisions to meet precise market demands.
The data has been captured from TOMRA machines worldwide and labeled by the company’s data science team, according to the release. The new-generation machine was tested for 18 months in the U.S. and New Zealand and demonstrated improved performance compared to its predecessor.
The LUCAi upgrade package can benefit existing Spectrim X customers who grade apples, according to TOMRA. New ‘plug-and-play’ deep learning models can detect and classify apple defects such as splits and punctures across multiple varieties. LUCAi also allows customers to adjust the severity of grading parameters, previously done by experienced operators, according to the release.
One of the first adopters of this technology was Freshco, a leading New Zealand fruit and vegetable grower and exporter.
“Spectrim with LUCAi is a game changer,” Robin Mudgeway, Frescho’s manager of technology and machinery, said in the release. “We had to work a lot harder around the stem area, mainly for splits and punches. Now you don’t have to do the work anymore, LUCAi does it for you. This means you can keep production up at a speed without a lot of manual grading on the line.”
TOMRA’s next LUCAi release will focus on cherries before the company turns to avocados, cherries, citrus, kiwifruit and stone fruit, according to the release.