A pose estimation approach for discarded stacked smartphones recycling: Based on instance segmentation and point...

Area of Science:

• Robotics and Automation
• Computer Vision
• Environmental Engineering

Background:

• The growing volume of end-of-life smartphones presents significant recycling challenges.
• Current manual disassembly processes are inefficient, labor-intensive, and environmentally problematic.
• Complex scenarios like stacking and occlusion hinder automated recycling of discarded smartphones.

Purpose of the Study:

• To develop an automated pose estimation method for stacked, discarded smartphones.
• To enhance the precision and efficiency of robotic grasping in e-waste recycling.
• To reduce manual intervention and environmental impact in smartphone disassembly.

Main Methods:

• Integration of an improved Mask R-CNN instance segmentation model with Iterative Closest Point (ICP) point cloud registration.
• Accurate segmentation of stacked smartphones using combined real and synthetic datasets.
• Development of a pose recognition interactive system for data visualization and dynamic interaction.

Main Results:

• Effective segmentation of stacked smartphones was achieved using the proposed method.
• Accurate pose information was extracted to guide robotic grasping actions.
• Transfer learning utilizing synthetic and real-world data demonstrated significant effectiveness.

Conclusions:

• The proposed pose estimation method significantly improves automation and efficiency in end-of-life smartphone recycling.
• The integration of advanced computer vision and point cloud registration offers a viable solution for complex disassembly tasks.
• This research provides valuable insights and technical solutions for intelligent e-waste management.