Machine Vision-Based Garbage Sorting System Demonstrates High Accuracy in Experimental Evaluations

Strategic Insights -

A team of researchers at the Wroclaw University of Science & Technology in Poland has developed a machine vision-based garbage sorting system that shows promise in efficiently and accurately sorting waste. The system, which integrates a robotic arm with advanced machine vision technologies, has been tested in various scenarios, including sorting individual garbage items, multiple non-adhering waste items, and adhesive or stacked waste items. According to the study, the system achieved a sorting accuracy of 90% for individual garbage items and 91.2% overall classification accuracy.

Key Takeaways:

  • The machine vision-based garbage sorting system integrates a robotic arm with advanced machine vision technologies to enhance sorting performance.
  • The system evaluates various network architecture models and develops a neural network using MobileNetV3 and YOLOv4 (You Only Look Once version 4) by optimizing the backbone network components.
  • The optimal grasp of the mechanical claws is determined using a method and minimum external torque algorithm, enabling the robotic arm to autonomously execute garbage sorting and classification tasks.
  • Experimental results demonstrate that the system achieves a sorting accuracy of 90% for individual garbage items and 91.2% overall classification accuracy.
  • In scenarios involving multiple, non-adhering waste items, the system consistently maintains a 100% sorting rate.
  • The neural network's target detection capabilities remain operational even when dealing with adhesive or stacked waste items, although the accuracy of sorting success may decrease.
  • The research validates the feasibility and reliability of the machine vision-based garbage sorting system through comprehensive experimental evaluations.
  • The system is designed to address challenges such as low efficiency, slow processing speed, and potential harm to human operators.

Statistics:

  • Sorting accuracy: 90% for individual garbage items
  • Overall classification accuracy: 91.2%
  • Sorting rate: 100% for multiple, non-adhering waste items
  • Target detection capabilities: neural network remains operational, although accuracy of sorting success may decrease

Sources:

  • A New Machine Vision-based Industrial Robot System for Garbage Management. Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, 2025.
  • Wroclaw University of Science & Technology
  • Sumika Chauhan, Wroclaw University of Science & Technology
  • Yaosheng Hu, Zixuan Yang, Qing Qu, Binbin Huang, Zhixiong Li, and Govind Vashishtha, Wroclaw University of Science & Technology