Umic, Unveils SINVICUBE, a Hyperspectral AI-Based Multi-Item Waste Sorting System

Umic, a company specializing in hyperspectral AI sorting technology, has unveiled SINVICUBE, a hyperspectral AI-based multi-item waste sorting system. The company is led by CEO Kim Do-hoon, and its global brand is SINVIAI. SINVICUBE was first unveiled at the 2026 Korea ESG Green Expo, held from October 21 to 23 at COEX in Seoul.

SINVICUBE is designed to sort multiple types of waste with a single piece of equipment. That differs from conventional optical sorting methods, which required separate machines for each sorted item. The system handles a variety of materials by changing software settings, and the items it can sort include PET flakes, black plastic, shredded waste batteries, marine waste, textiles, glass, and EPS.

The system's main components are a hyperspectral camera, chute, and air ejection device, and it was designed with a variable structure that can be adjusted depending on the sorting target. The chute is a device that forms the material transfer path, and it was built so that optical conditions and drop paths can be adjusted according to the characteristics of each item. AI analyzes multiple image data sets in the ultraviolet (UV) to mid-infrared (Mid-IR) wavelength range to distinguish materials by composition, shape, and contamination level.

It is equipped with SDS, a sorting software platform independently developed by Umic. When the sorting target changes, the platform can set sorting policies suited to the item or add new ones, and it also supports policy changes and performance improvements through OTA updates. The platform is offered in two deployment modes: SDS-One for new equipment and SDS-Ready, which is connected to existing sorting machines.

At recycling sorting sites, operational considerations include not only configuring equipment by item, but also whether sorting conditions can be adjusted in response to changes in processed materials. Umic expanded its lineup of sorting equipment with the release of the new product. Its existing lineup covered raw-material sorters and shredded-material sorters, and after the expansion, it now includes multi-item sorters.

Umic said it has cumulatively installed 25 optical sorting machines at public and private sorting facilities nationwide. According to test results from the Korea Testing Laboratory (KTL), the separation accuracy was 98.4%. The company holds a total of 8 registered patents related to hyperspectral technology. Umic is working with research institutions to expand the scope of its sorting technology applications.

On October 27, Umic and the Korea Institute of Geoscience and Mineral Resources (KIGAM) held the launch ceremony for the EPS packaging recycling technology commercialization council. At the ceremony, the two sides discussed developing AI-based automated multi-stage sorting technology for waste EPS and technology to remove contaminants. Umic and the Korea Institute of Geoscience and Mineral Resources (KIGAM) plan to review the potential for field demonstrations and commercialization, and to pursue planning for a government-supported R&BD project in 2027.

Kim Do-hoon, CEO of Umic, said SINVICUBE was developed not as a replacement for equipment, but on the basis of a software expansion approach, and that its purpose is to respond to diverse sorting needs. He added that the company plans to continue advancing its sorting technology through cooperation with research institutions and expand its application across recycling operations.

Source: TECHWORLD · Kim Gyeong-ju
Original: https://www.epnc.co.kr/news/articleView.html?idxno=409303

References

• https://www.epnc.co.kr/news/articleView.html?idxno=409303