Yachiyo Engineering Co., Ltd.
Menu
Menu
Overview
In collaboration with Ehime University, we have jointly developed "PRIMOS," a new AI-based system that automatically detects and classifies plastic waste from videos of river surfaces, which we began offering on April 14, 2025.
This system solves challenges that conventional methods struggled with, such as identifying waste types, reducing labor and costs, and reliance on observer skills, enabling more accurate and objective monitoring. The aim of PRIMOS is to contribute to a fundamental solution to the marine plastic waste problem and to the protection of the natural environment.
Visit this website for more details
In collaboration with Ehime University, we have jointly developed "PRIMOS," a new AI-based system that automatically detects and classifies plastic waste from videos of river surfaces. We began offering this service on April 14, 2025.
This system solves challenges that conventional methods struggled with, such as identifying waste types, reducing labor and costs, and reliance on observer skills, enabling more accurate and objective monitoring. The aim of PRIMOS is to contribute to a fundamental solution to the marine plastic waste problem and to the protection of the natural environment.
By analyzing video data of rivers, the system identifies waste floating on the river surface, detects plastic waste with AI (YOLOv8 instance segmentation model), and classifies it into four categories: "PET bottles," "food containers," "plastic bags," and "other plastics." In demonstration tests, this method has shown high detection accuracy.
The AI can also detect the flow speed at the river surface, allowing it to calculate how much plastic waste is being carried downstream. Using the relationship between the actual area and weight of plastic waste, the system is able to determine the transport volume per unit of time from the waste's converted weight and the water flow rate calculated from its speed and the river's cross-sectional area. By taking into account factors like population distribution and land use, the system can estimate the total amount of waste flowing into rivers from across the entire basin. This information can then be used to support local environmental conservation activities.
Large amounts of plastic waste in rivers often appear when the flow rate increases during rainfall. However, the system is flexible enough to handle these changes, providing stable detection performance even in videos taken during high-water floods.
Surveys that use installed cameras to film video can significantly reduce the time, labor, and cost associated with manual surveys. This method also allows for safe data collection during floods when a large amount of debris is present. In addition, the data can be collected and analyzed based on a unified standard, so it does not depend on the observer's skill. This means the information can be recorded as objective data. Furthermore, image data can be centrally managed and permanently stored, which enables a more accurate understanding of the situation.
Conventional monitoring method
Monitoring Method Using PRIMOS