验证了聚类分析用于卫星雷达测高波形分类的可行性。针对雷达测高波形的特征,提出了移动最小欧氏距离算法,并以该距离作为相似性统计量进行聚类分析,取得了很好的效果。以Topex/Poseidon卫星的近岸波形为例,比较了最短距离法、最长距离法、类平均法和最小方差法等4种聚类方法的聚类效果,结果表明最小方差法效果最优。
The feasibility of classifying satellite radar altimeter waveforms by cluster analysis is validated.According to the characteristic of radar waveforms,moving minimum Euclidean distance is proposed to be the similarity index for cluster analysis.This method can efficiently classify the radar waveforms.Taking the Topex/Poseidon coastal waveforms as an example,we compare four cluster analysis methods,such as single linkage,complete linkage,average linkage and Ward linkage.Results show that the Ward linkage cluster analysis is better than the others.