遙感數據處理中多源數據融合方法研究

打開文本圖片集

中圖分類號：P237 文獻標志碼：A 文章編號：2095-2945(2025)14-0162-04

1,2,3（1.,西安710001；2.陜西省第三測繪工程院,西安71001；3.西安航空學院,西安 710089）

Abstract:Inremotesensing dataprocesing，multi-sourcedata fusion method hasbecomea keytechnologytoimprove the eficiencyandaccuracyof informationextraction.Thispapersystematicalldiscussesthemulti-sourcedatafusionmethodbased onfeaturespaceanddeeplearning,analyzesthediferencebetweenlinearandnonlinearfusionanditsapplicationintheproce offeaturextractionandselection,andfurtherelaboratestheimportanceoffeaturespacedimensionreductiontechnology.Dep learningtechniques，speciallconvolutionalneuralnetworksandself-supervisedlearning，havedemonstratedexcellent performanceinprocessngheterogeneousandmulti-dimensionalremotesensingdata，signficantlyimprovingtheaccuracyand robustnessofdatafusion.Basedonpracticalcases,thispapershowsthespecificefectsofdiferentfusionmethodsonfeature extractionandfusioninmulti-sourcedataprocessng，indicatingthatdeeplearning methodshavebroadapplicationprospectsin the field of remote sensing.

Keywords: remote sensing data; multi-source data fusion; feature space; fusion effect; deep learning

在遙感技術的不斷進步與廣泛應用背景下，多源數據融合已成為提升數據處理精度與豐富信息內容的關鍵手段，遙感數據來源廣泛，涵蓋光學、雷達、激光雷達等多種傳感器，這些數據在空間、時間及光譜分辨率上各具特性。（剩余5629字）