Methods, devices and equipment for detecting defects in bolts of power facilities

By employing a high-frequency detail-guided feature enhancement module, a decoupled context and saliency pyramid pooling module, and an asymmetric guided fusion pyramid network, the problems of lost detail information and semantic conflicts in the detection of bolts in power facilities are solved, thereby improving detection accuracy and stability.

CN122312486APending Publication Date: 2026-06-30STATE GRID HEBEI ELECTRIC POWER CO LTD BAODING POWER SUPPLY BRANCH CO +2
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
STATE GRID HEBEI ELECTRIC POWER CO LTD BAODING POWER SUPPLY BRANCH CO
Filing Date
2026-02-09
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

Existing methods for detecting bolt defects in power facilities suffer from problems such as loss of detailed information, insufficient utilization of context, and semantic conflicts in multi-scale feature fusion, leading to high misjudgment rates and insufficient positioning accuracy.

Method used

We employ a high-frequency detail-guided feature enhancement module, a decoupled context and saliency pyramid pooling module, and an asymmetric guided fusion pyramid network to improve detection accuracy through high-frequency detail extraction, complementarity of saliency and context information, and cross-scale feature fusion.

Benefits of technology

It improves the detection rate and positioning stability of micro-bolt defects in complex inspection scenarios, reduces the false judgment rate, and improves the detection accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122312486A_ABST
    Figure CN122312486A_ABST
Patent Text Reader

Abstract

This invention provides a method, apparatus, and equipment for detecting bolt defects in power facilities, relating to the field of power technology. The method includes: acquiring inspection images of bolts in power facilities; downsampling the inspection images to extract original features; inputting the original features into a feature extraction network constructed by a high-frequency detail-guided feature enhancement module and a decoupled context and saliency pyramid pooling module to extract multi-scale features; performing cross-scale feature fusion on the multi-scale features through an asymmetric guided fusion pyramid network constructed by the high-frequency detail-guided feature enhancement module, an asymmetric path splitting module, and a detail-guided fusion module to obtain fused features; and processing the fused features through a detection head to obtain the bolt defect detection result. This invention can solve the problems of lost detail information, insufficient utilization of context, and semantic conflicts in multi-scale feature fusion that exist in existing methods when detecting bolts.
Need to check novelty before this filing date? Find Prior Art