对伶仃洋表层沉积物的聚类分析

Lingdingyang surface sediments clustering analysis based on neural network

本文采用模糊等价聚类、模糊C-均值聚类法对伶仃洋河口175个沉积物样本做了聚类分析,讨论了两种方法对于沉积物聚类的适用性和各类沉积物在伶仃洋河口的分布区域,结果表明:对沉积物的聚类分析模糊C-均值聚类法较模糊等价聚类法更为适用,聚类结果表明:第一类沉积物主要分布与内伶仃岛北侧、东侧以及向大濠延伸的区域,其余研究区域有少量分布。第二类沉积物分布区域最为广泛,主要分布于虎门口外到内伶仃岛的整个偏东区域、内伶仃岛东南方以及淇澳岛南侧,且以黏土质粉砂为主。第三类沉积物主要分布在洪奇门、横门口外东南侧,粒径较粗以砂为主,第四类沉积物较少,各个区域有少量分布。

In this paper,two fuzzy clustering method to analyze the Ling Ding 175 sediment samples,both methods discussed clustering applicability for the sediment deposits and various distribution areas in the estuary Ling Ding,the results showed:Fuzzy C-means clustering method is more fuzzy equivalent clustering is more suitable for clustering analysis of sediment,Clustering results showed that sediments mainly with first class Neilingding island north of the east and the region extending to Da hao island,a small amount of the distribution of the rest of the study area.The second category is mainly distributed in Humen estuary to the entire region east Neilingding Island,Neilingding island south-east,Qiao Island on the south side.The third type deposits are mainly distributed in Hongqi door,coarse particle size to sand-based,less sediment fourth category,each region has a small amount of the distribution.