Drop-out fuse state monitoring method based on multi-size features and deep learning

By improving the YOLOx network and combining multi-scale feature decoupling and dynamic fusion methods, the problem of insufficient accuracy and efficiency in drop-out fuse detection in the power industry is solved, achieving efficient and accurate small target detection, which is suitable for power drone platform applications.

CN121963075APending Publication Date: 2026-05-01PUYANG POWER SUPPLY COMPANY STATE GRID HENAN ELECTRIC POWER
0 Cites 0 Cited by

Patent Information

Application Number
CN202511897213.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-16
Publication Date
2026-05-01

AI Technical Summary

Technical Problem

Existing monitoring models in the power industry lack sufficient accuracy and efficiency in detecting drop-out fuses, especially in terms of the accuracy and reliability of detecting small targets in complex backgrounds, which needs to be improved, and also consume a lot of hardware resources.

Method used

An image quality optimization system based on multi-scale feature decoupling and dynamic fusion is constructed. By improving the YOLOx network, receptive field blocks and coordinate attention modules are added. Combined with adaptive spatial feature fusion module and lightweight improvements, a state monitoring model is constructed. Data augmentation and annotation are performed using UAVs to acquire multi-angle images. A loss function that combines weighted loss and localization loss is introduced to optimize model performance.

Benefits of technology

It improves the accuracy and efficiency of drop fuse detection, reduces model complexity, makes it suitable for embedded platforms of power drones, reduces the inspection burden, and improves detection effect and identification accuracy.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a drop-out fuse state monitoring method based on multi-size features and deep learning, and the method comprises the steps: S1, forming a data set for obtained drop-out fuse pictures, and dividing the data set into a training set and a test set; s2, adding an improved receptive field block and a coordinate attention module to a YOLOx backbone network, adding an adaptive spatial feature fusion module to PANet, carrying out secondary fusion on features of different scales, then introducing a loss function of weighting loss and positioning loss fusion, and finally carrying out lightweight improvement, constructing a state monitoring model, and carrying out state monitoring. Performing training verification on the constructed state monitoring model through the training set and the test set; s3, identifying a drop-out fuse picture acquired in real time by adopting the trained and verified state monitoring model; according to the image data collected by the application, the construction of a special database is realized, the YOLOx is improved and lightweight operation is carried out, and the complexity of the model is reduced, so that the method is suitable for an embedded platform of an electric unmanned aerial vehicle, and the detection effect of the drop-out fuse is improved.
Need to check novelty before this filing date? Find Prior Art