/recyclebot-project

recyclebot-project

Expected Deliverables

  1. Dataset containing:
    • RGB images with plastic packaging items as classes (such as bottles, bags, trays, pots, and tubs).
    • Hyperspectral images with seven plastic types as classes: PET, PE-HD, PVC, PE-LD, PP, PS, O.
  2. Demonstration setup. Consists of:
    • Sensors.
      • RGB camera.
      • Hyperspectral camera.
    • Manipulator.
      • Delta, articulated, or SCARA.
    • Gripper.
      • Vacuum suction (single cup or an array with level compensation).
    • Conveyor belt

Demonstrator

Expected functional flow:

  • A bulk of plastic waste moves on the conveyor belt passing hyperspectral and RGB cameras towards the robot.
  • The object detection model detects individual plastic items.
  • The classification model classifies each plastic item.
  • The command server creates a pick-and-place task for the robot based on the detection and classification models' outcomes.
  • The robot executes the pick and place task and moves plastic items from the belt to relevant bins.

Recyclebot Demonstrator