Line-Features-Based Pose Estimation Method for the Disc Cutter Holder of Shield Machine

Area of Science:

• Robotics
• Machine Vision
• Mechanical Engineering

Background:

• Automatic disc cutter replacement in shield machines requires precise pose estimation of the disc cutter holder.
• Polluted and low-illumination environments pose significant challenges for traditional machine vision methods.

Purpose of the Study:

• To develop a robust line-feature-based pose estimation method for shield machine disc cutter holders using a monocular camera.
• To address challenges posed by blurred edges and environmental conditions in industrial settings.

Main Methods:

• A rounded edge model was developed to accurately detect edge points, matching the 3D model of the workpiece.
• An adaptive threshold region growing method was employed for contour separation and edge search box generation.
• Preprocessing of edge points and RANSAC linear fitting were utilized to improve efficiency and accuracy.

Main Results:

• The proposed method demonstrated high reliability in pose estimation for disc cutter holders.
• The technique effectively handles blurred edges caused by rounded corners.
• Experimental results confirmed the method's suitability for practical engineering applications.

Conclusions:

• The developed line-feature-based pose estimation method significantly improves the accuracy of disc cutter holder localization.
• This approach is crucial for enabling reliable automatic disc cutter replacement in shield machines.
• The method meets the stringent measurement accuracy requirements of real-world engineering applications.