一种车载接触网异常图像智能识别方法及装置
By collecting and storing the catenary image sequence and performing feature recognition, combined with in-frame and out-of-frame verification, the problem of image quality fluctuation in the vehicle-mounted catenary was solved, enabling accurate identification and early warning of catenary anomalies and reducing false alarm and missed detection rates.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 成都骏盛科技有限责任公司
- Filing Date
- 2026-04-29
- Publication Date
- 2026-07-17
AI Technical Summary
Existing vehicle-mounted overhead contact line images exhibit significant quality fluctuations and rely on single-frame image analysis, making it difficult to effectively distinguish between real defects and transient noise. This results in low detection efficiency, high labor intensity, and difficulty in timely detection and accurate location of minute defects.
The original image sequence of the overhead contact line is acquired and associated coordinate parameters are obtained and stored. Complex background interference is suppressed by frame-by-frame matching and alignment and background removal. A dual-branch feature extraction model is used to identify texture features and structure. Combined with intra-frame verification and inter-frame linkage verification, a comprehensive recognition result is obtained.
It enables accurate identification and early warning of abnormal conditions in the overhead contact system, reduces false alarm and missed detection rates, and improves the accuracy of target area extraction and the robustness of detection.
Smart Images
Figure CN122115457B_ABST