Fittings Detection Method Based on Multi-Scale Geometric Transformation and Attention-Masking Mechanism

Area of Science:

• Electrical Engineering
• Computer Vision
• Artificial Intelligence

Background:

• Overhead transmission lines are critical infrastructure for power systems.
• Intelligent patrol technology is essential for smart grid development.
• Current fittings detection methods struggle with variations in scale and geometry.

Purpose of the Study:

• To develop an advanced fittings detection method for overhead transmission lines.
• To enhance the accuracy and efficiency of intelligent patrol systems.
• To address limitations in detecting fittings with diverse scales and geometric properties.

Main Methods:

• A multi-view geometric transformation enhancement strategy using homomorphic images.
• An efficient multiscale feature fusion technique for improved target detection.
• An attention-masking mechanism to optimize computational load and feature learning.

Main Results:

• The proposed method significantly enhances the detection accuracy of transmission line fittings.
• Experimental results demonstrate superior performance across various datasets.
• The integration of multi-scale features and attention mechanisms proves effective.

Conclusions:

• The developed fittings detection method offers a substantial improvement for intelligent power system monitoring.
• This research contributes to the advancement of smart grid technologies.
• The method effectively handles scale variations and geometric complexities in fitting detection.