基于视觉原语的电力边缘异构融合放电识别方法及系统
By constructing a heterogeneous fusion discharge recognition method for power edge based on visual primitives, and combining visual primitive data oriented towards cognition and motion, efficient and accurate recognition of partial discharge in GIS is achieved. This solves the recognition difficulties of traditional image sensors in complex environments, improves detection accuracy and real-time performance, and supports intelligent management of GIS equipment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA ELECTRIC POWER RESEARCH INSTITUTE CO LTD
- Filing Date
- 2025-05-14
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional image sensors struggle to effectively identify partial discharges in gas-insulated metal-enclosed switchgear (GIS) when faced with complex and ever-changing anomalies, leading to algorithm failure.
A heterogeneous fusion discharge identification method based on visual primitives is adopted. By collecting visual primitive data oriented towards cognition and motion, a target detection model is constructed. Combining convolutional modules, connection layers, upsampling layers, pooling layers and detection modules, the location and category of partial discharge in GIS are detected, and high-quality color images are generated through visual reconstruction model.
It enables efficient and accurate identification of partial discharge in GIS, improves detection accuracy and speed, ensures the real-time nature and interpretability of results, and supports intelligent management of GIS equipment.
Smart Images
Figure CN120543923B_ABST
Abstract
Citation Information
Patent Citations
Electric power operator identification method and system based on deep learning and binocular vision
CN118071983A
GIS equipment partial discharge mode identification method and system, terminal equipment and storage medium
CN119805118A