Method for judging the attitude of scissor type isolation switch

This paper proposes a method of using machine vision technology related to deep learning to automatically determine the posture of scissor type isolation switches based on photos of the operation area of the power grid where the scissor type isolation switch is located. Firstly, by constructing an image dataset of scissor type isolation switches, opening and closing posture style images were selected, and feature vectors describing landmarks and features were extracted. Next, by using the method of marker point pairing, the pairing points are selected to determine the scissor style isolation switch posture of the image to be paired. The innovation of this paper lies in proposing a new method for generating descriptive feature vectors and pairing feature points. By establishing the scale space of the image and extracting features, it can more accurately determine the posture of the scissor type isolation switch, improving the efficiency and reliability of automatic recognition. This research achievement is of great significance in power grid operation and maintenance, which can improve the safety and efficiency of power grid equipment maintenance, and contribute new technological means and methods to the development of the power industry.