Power transmission channel fire point identification method based on multi-scale spatial-temporal characteristics

CN121617015APending Publication Date: 2026-03-06YANBIAN ELECTRICAL BUREAU +1
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202511836343.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-08
Publication Date
2026-03-06

AI Technical Summary

Technical Problem

Existing methods for detecting fire points in power transmission channels are mostly based on static identification using single-frame infrared or visible light images. They lack analysis of the changing patterns of fire points over time, leading to false detections or missed detections. Furthermore, traditional methods are difficult to adapt to diverse environments and lack the ability to distinguish between deep semantic and spectral features.

Method used

A fire point identification method based on multi-scale spatiotemporal features is adopted. Through preprocessing of multi-modal monitoring data, multi-scale adaptive spatiotemporal encoder, dynamic evolution prediction model and physical simulation consistency judgment, combined with adaptive fusion and online self-learning, dynamic analysis and interference elimination of fire points are achieved.

Benefits of technology

It improves the robustness of fire point identification, reduces false detections and false negatives, achieves accurate identification of real fire points, and has self-learning and environmental self-adaptation capabilities, adapting to diverse power transmission channel environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121617015A_ABST
    Figure CN121617015A_ABST
Patent Text Reader

Abstract

The invention provides a power transmission channel fire point identification method based on multi-scale spatial-temporal characteristics, and relates to the technical field of power transmission channels, and the method comprises the steps: firstly collecting multi-modal monitoring data, carrying out the resolution enhancement and variable time window coding of a suspected fire point through a multi-scale adaptive spatial-temporal encoder, and generating a multi-scale spatial-temporal characteristic spectrum; and inputting the characteristics into a fire point dynamic evolution prediction model and a typical interference dynamic evolution prediction model, comparing prediction errors, eliminating interference which is not consistent with a dynamic change rule, back-projecting the fire point to a three-dimensional scene model, and running simplified thermal plume or smoke plume diffusion simulation to obtain a three-dimensional scene model. The expected physical effect is compared with the actually observed dynamic characteristics, the optical interference which does not conform to the thermodynamic law is thoroughly eliminated, finally, the comprehensive consistency confidence coefficient is calculated, high-sensitivity recognition of early-stage small fire points is achieved, meanwhile, the method has extremely high robustness to false fire points such as sunlight reflection and environmental change, and the method is suitable for large-scale popularization and application. The method is suitable for complex and diverse power transmission channel environments.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Cited By

  • Intelligent analysis method and system based on multi-modal data fusion

    CN121858396A

  • Emergency fire-fighting event identification method and system for industrial enterprises

    CN122020262A

  • An emergency fire fighting event identification method and system for industrial enterprises

    CN122020262B

  • Remote sensing image enhancement method for forest fire monitoring

    CN122115252A