Fire point monitoring data processing method and device based on optical remote sensing image data

By combining preprocessing of optical remote sensing image data with deep learning and spectral hybrid decomposition, the problem of insufficient accuracy and complex background recognition in existing fire detection methods has been solved, achieving high-precision and rapid fire detection that is adaptable to diverse global environments.

CN122454427APending Publication Date: 2026-07-24CHINA THREE GORGES CORPORATION
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610453689.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-08
Publication Date
2026-07-24

AI Technical Summary

Technical Problem

Existing fire detection methods, when accuracy requirements increase, suffer from problems such as missed detection of weak fire points, high false alarm rate, insufficient ability to identify complex backgrounds, and complex operational processes requiring extensive manual intervention.

Method used

A fire detection method based on optical remote sensing image data is adopted. Through preprocessing, multidimensional feature extraction, integrated machine learning and deep convolutional neural network combined with attention mechanism, fire pixel identification is performed, and spectral mixing decomposition is carried out. Combined with cloud and smoke interference correction, high-precision fire detection is achieved.

Benefits of technology

It improves the accuracy of fire point identification, reduces the false alarm and missed detection rates, meets the minute-level response requirements, adapts to complex geographical environments and combustion conditions, and reduces human intervention.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122454427A_ABST
    Figure CN122454427A_ABST
Patent Text Reader

Abstract

Embodiments of the present disclosure relate to a fire point monitoring data processing method and device based on optical remote sensing image data, wherein the method comprises: preprocessing an original satellite image to obtain a to-be-processed satellite image; extracting a plurality of features of each pixel in the to-be-processed satellite image to construct a multi-dimensional feature vector, and performing initial screening based on the multi-dimensional feature vector to obtain a potential fire point pixel; identifying the potential fire point pixel to output a probability value of the potential fire point pixel being a fire point, and determining that the potential fire point pixel is a suspected fire point pixel when the probability value is greater than a first preset threshold; obtaining the abundance of the suspected fire point pixel, and confirming that the suspected fire point pixel is a real fire point pixel when the abundance is higher than a second preset threshold, and outputting the geographical coordinates, the burning intensity and the confidence parameter corresponding to the real fire point pixel. The above technical solution has stronger recognition ability for small and weak fire points and fire points in a singular background, and greatly reduces the false positive rate of fire points.
Need to check novelty before this filing date? Find Prior Art