A Training-Free Foreground-Background Separation-Based Wire Extraction Method for Large-Format Transmission Line...

Area of Science:

• Electrical Engineering
• Computer Vision
• Image Processing

Background:

• Smart grids require advanced monitoring of transmission lines.
• Accurate wire extraction is essential for defect detection in large-format images.
• Existing methods often require extensive training data and struggle with complex backgrounds.

Purpose of the Study:

• To develop a training-free wire extraction method for large-scale transmission line images.
• To enhance the separation of wires from complex backgrounds using depth estimation.
• To provide an efficient and dataset-independent solution for wire identification.

Main Methods:

• Leveraging depth estimation maps to improve foreground wire and background separation.
• Employing a line segment structure-based method to identify horizontally oriented linear features.
• Developing a training-free and dataset-independent approach.

Main Results:

• Effective separation of transmission wires from complex backgrounds.
• Robust identification of slender wire structures in cluttered scenes.
• Reduced computational overhead compared to deep learning methods.

Conclusions:

• The proposed training-free method accurately extracts transmission wires from challenging images.
• Depth estimation significantly enhances wire-background separability.
• This approach offers an efficient alternative for smart grid transmission line monitoring.