Page 205 - AI for Good Innovate for Impact
P. 205
AI for Good Innovate for Impact
Use Case 6: AI-powered Precision Recycling System (WIMPLE)
Organization: Seoreu. Co., Ltd. Change 4.2-Climate
Country: SOUTH KOREA
Contact Person(s):
Primary: Jonghyuck LEE, jhlee@ seoreu .com
Secondary: Yeoun Joo LEE, yjlee@ seoreu .com
1 Use Case Summary Table
Item Details
Category Climate Change/Natural Disaster
Problem Addressed Once recyclable waste is mixed with general waste, it becomes extremely
difficult to recycle it due to the significant costs involved in separating
recyclable materials from mixed waste. To improve these rates, sorting
at the source is essential and must be done with high accuracy and
precision.
The main reason for poor sorting practices lies in the complexity of
recycling guidelines, which vary by country and even by region—and
frequently change. High-precision sorting, essential for resource circu-
lation-the ultimate goal of recycling, requires active participation from
all citizens, which is virtually impossible. As a result, high-cost systems
like deposit-refund schemes are being adopted to encourage partic-
ipation and increase recycling rates. However, it not only covers part
of the recycling waste but also costs too much to maintain the system.
Key Aspects of Solu- Most recyclable materials generated in daily life are disposable contain-
tion ers. To achieve accurate sorting, it is first necessary to classify these
containers by material type—paper, plastic, metal, glass, etc. The purer
the material grouping, the higher its value as a recyclable resource.
Further classification is also needed, e.g., aluminum vs. steel cans, card-
board vs. carton paper. Additionally, it must be determined whether each
item is recyclable and whether additional processing is required.
AI enables this granular classification and communicates the sorting
information to various robotic systems to ensure precise and automated
separation. The system can adapt to different national standards and be
applied to deposit-refund schemes.
Technology Keywords Object detection, feature matching, object recognition
Data Availability Public (developed by Seoreu Co, Ltd) dataset: https:// www .aihub .or .kr/
aihubdata/ data/ view .do ?currMenu = 115 & topMenu = 100 & aihubDataSe
= data & dataSetSn = 71385 [4]
Metadata (Type of Visual (spatial data)
Data)
169
