On the 11th, the unloading area of DaehanSteel's Noksan Recycling Center (NRC) in Gangseo District, Busan. As a truck loaded with scrap metal entered the storage area, a camera mounted on the ceiling moved to match the truck's position. Two cameras equipped with the AI-based scrap inspection solution, AIMOS, automatically recorded the moment when scrap metal was lowered by a claw.
In the control room within the storage area, the type and grade of the scrap metal brought in by the truck, as well as any hazardous materials, were verified in real-time through the camera. This information is also sent to the customer company that sent the scrap.
Scrap metal, along with iron ore and coke, is one of the three main raw materials in the steel industry. It is melted into molten iron in electric furnaces and reborn as products such as automobiles and machinery. Since various types are mixed together during distribution, sorting is crucial.
Currently, most grading work is performed by inspectors' eyes. Since inspection work takes place simultaneously with unloading in the storage area where scrap metal is piled up, meticulous sorting is difficult. If a specific item appears to be more than 70% of what the inspector sees, the entire truck may be classified as that item. Scrap metal loaded in a 20-ton truck varies greatly in value, ranging from 8 million won to over 20 million won, depending on the item and grade.
AIMOS is a service created by DaehanSteel and LG CNS to replace inspection work that relied on human eyes with AI. It is also the name of the joint venture established by DaehanSteel and LG CNS in a 70-30 ratio.
AIMOS has learned from 2 million scrap metal images to determine grades more accurately than humans and quickly identify hazardous materials. The time for unloading work is between 5 to 10 minutes, and inspectors lack sufficient time to thoroughly see.
Bom Kim, CEO of AIMOS, noted, "Depending on the storage environment, as the assessment through AIMOS becomes more advanced, one inspector can manage up to five storage locations," adding, "Existing inspectors can manage scrap data while performing other tasks, increasing efficiency."
At the DaehanSteel NRC visited that day, there were three storage areas capable of holding 25,000 tons of scrap metal on a site of 8,500 square meters (about 2,570 pyeong), but there was only one employee sitting in the control room watching the monitors. The storage area had only the minimum personnel needed to guide the unloading work of trucks and prepare for safety accidents.
Kim said, "If AI assessments become widespread, it can reduce safety incidents, and the transparency of price calculations can increase, thereby innovating distribution methods."
AIMOS began a pilot operation last year at DaehanSteel and YK Steel, and from early this year, it has been introduced to SP Nature, one of the large scrap distribution companies in South Korea, in the form of Software as a Service (SaaS). It is scheduled to be installed at six additional scrap distribution companies by the third quarter of this year. There are a total of nine steelmakers affiliated with the Korea Steel Association's Scrap Committee, including DaehanSteel and YK Steel, and about 200 distribution companies.
AIMOS anticipates an increase in demand for scrap management as the steel industry accelerates its transition to electric furnaces. Electric furnaces make molten iron from scrap metal. Kim said, "With the prototype released this year, we plan to gradually expand our presence."