简介:Identificationofforestfire_pointsinNOAAimagesisthebasisofmonitoringforestfireusingNOAAsatellitedata.Traditionalvisualinterpretationisdifficulttosettleforauto_identificationwithcomputer.Theartificialneuralnetworktechniqueprovidesanewmeansforsolvingthisproblem.Inthispaper,theprinciplesandmethodofusingneuralnetworktechniquetoautomaticallyidentifyfire_pointsinNOAAimagesarediscussedandthetestintherangeofHubeiprovinceispresented.Theresultofthetestshowsthatthedisciplinedneuralnetworkhascollectedthecharacteroffire_pointsandhasabilitytoidentifyfire_pointsinNOAAimages.Comparingneuralnetworkwithvisualinterpretation,theconclusionisdrawnthatbyusingneuralnetworkthepurposeofauto_identificationofforestfire_pointsinNOAAimagescanberealizedwiththealmostsameprecision.