基于视觉干扰抑制的古建筑早期火焰图像识别方法及系统
By constructing a physical evolution model of smoke and flame in the early stage of a fire and conducting spatiotemporal consistency analysis, extracting spatiotemporal consistency feature sequences and suppressing visual interference, the problem of distinguishing between flames and smoke in ancient buildings was solved, and accurate identification of early flame images was achieved.
Patent Information
- Application Number
- CN202610044855.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-14
- Publication Date
- 2026-07-17
- Estimated Expiration
- 2046-01-14
AI Technical Summary
In ancient buildings, the smoke from burning smoke is similar to that of early flames in terms of diffusion patterns, brightness changes, and motion characteristics, resulting in insufficient accuracy in flame image recognition and making it difficult to accurately distinguish between the smoke and flames.
By constructing a physical evolution model of smoke and flame in the early stage of a fire, spatiotemporal consistency analysis is performed, spatiotemporal consistency feature sequences are extracted, visual interference suppression processing is carried out, candidate flame regions are screened, and judgment and identification are made in combination with smoke concentration and temperature information.
It improves the accuracy and stability of early flame image recognition in ancient buildings and solves the problem of flame recognition under smoke interference.
Smart Images

Figure CN121884245B_ABST
Abstract
Citation Information
Patent Citations
Intelligent position service system based on Beidou navigation
CN118396335A
Self-adaptive composite fire detection alarm model construction method for historic building
CN119418457A
Fire early warning and intelligent fire extinguishing method based on image recognition
CN120726749A
Camera-based flame recognition method and device, electronic equipment and storage medium
CN121074436A