高海况条件下海面漂浮小目标特征提取与分析

在高海况条件下,海面小目标回波信杂比很低,且在海浪影响下,小目标姿态变化引起的能量起伏严重,以能量为主导的目标检测方法具有较大的性能损失。从不同表示域提取小目标与海杂波差异特征,并进行特征检测是解决该问题的有效途径。以此为背景,通过开展海面小目标数据采集试验,分别从时域、频域、时频域、时频脊变换域进行特征提取与可分性分析,在 4个表示域共提取出 10个特征,包括时域相对平均幅度(RelativeAverageAm-plitude,RAA)、频域相对多普勒向量熵(RelativeVectorEntropy,RVE)、时频脊累积量(RidgeIntegration,RI)等,通过分析得到了 4级海况条件下不同特征可分性程度的结论,为后续多特征融合以及多维特征目标检测工作提供支撑。

FeatureExtractionandAnalysisofSmallFloatingTargets inHighSeaConditions

Undertheconditionofhigh seastate,theechosignal toclutterratioofsmall targets ontheseasurfaceisvery low, andthe energyfluctuationcausedbythechangeofattitudeofsmall targetsundertheinfluenceof wavesisserious, whichleads tothegreatperformancelossof energy-oriented targetdetectionmethods.Extractingthedifferencefeatures of small targets and sea clutter from different representation domains and detecting the features is an effective way to solvethisproblem.Based onthisbackground,asmalltargetdataacquisitionexperimentontheseasurfaceiscarriedout, feature extraction and divisibility analysis are carried out from the time domain, frequency domain, time-frequency do-mainandtime-frequencyridgetransformdomainrespectively,andatotalof10featuresareextractedfromthefourrepre-sentationdomains,includingtherelativeaverageamplitudeinthetimedomain,relativeDopplervectorentropyinthefre-quencydomain,andtime-frequencyridgeintegration, etc.Through theanalysis,theconclusionofthedivisibilitydegree ofdifferentfeaturesundertheconditionofclass4seastateisobtained,whichlays afoundati on for hesubsequentmulti-dimensionalfeaturefusionandmulti-featuretargetdetection.