图像火检算法的性能评估与优化策略研究.pdfVIP

安全生产SAFEPRODUCTION

图像火检算法的性能商优化策略研究

叶红武1张林张治湖?宋继坤2

431614；2.烟台龙源电力技术股份有限公司，山东烟台2

1.国能长源汉川发电有限公司，湖北武汉265503

摘要：图像火检算法作为火灾预警与监测的关键技术，在复杂环境下的性能尤为重要。文章针对图像火检算法在复杂环境下的

性能进行评估，发现强光、火焰形状变化影响其准确率。为改善此问题，提出预处理优化、特征精确提取及算法融合与集成学

习三大策略。实施后，算法在复杂条件下的准确率得到显著提升，验证了优化策略的有效性，为火检算法性能评估提供了系统

方法，优化策略对提升火灾预警与监测的准确性和鲁棒性具有重要价值。

关键词：图像火检算法；性能评估；预处理技术文献标识码：A中图分类号：TP391

D0l:10.13535/j.cnki.10-1507/n.2025.08.49

文章编号：2096-4137（2025）08-158-03

Researchonperformanceevaluationandoptimizationstrategiesofimagefiredetectionalgorithm

YEHongwu,ZHANGLin,ZHANGZhihu²,SONGJikun?

1.GuonengChangyuanHanchuanPowerGenerationCo.,Ltd.,Wuhan431614,China;2.YantaiLongyuanPowerGroupCo.,Ltd.,Yantai265503,

China

Abstract:Theimagefiredetectionalgorithm,asakeytechnologyforfirewarningandmonitoring,hasparticularlyimportant

performanceincomplexenvironments.Thisstudyevaluatestheperformanceofimagefiredetectionalgorithmsincomplex

environmentsandfindsthatstronglightandchangesinflameshapeaffecttheiraccuracy.Toimprovethisproblem,threestrategies

areproposed:preprocessingoptimization,precisefeatureextraction,algorithmfusion,andensemblelearning.Afterimplementation,

theaccuracyofthealgorithmwassignificantlyimprovedundercomplexconditions,verifyingtheeffectivenessoftheoptimization

strategy.Thisstudynotonlyprovidesasystematicmethodforevaluatingtheperformanceoffiredetectionalgorithms,butits

optimizationstrategyhasimportantvalueinimprovingtheaccuracyandrobustnessoffirewarningandmonitoring.

Keywords:imageflamedetectionalgorithm;performanceevaluation;preprocessingtechniques