傅里叶识别分析在赤潮生物自动识别中的应用研究  

FOURIER RECOGNITION ANALYSIS IN THE APPLICATION OF AUTOMATIC RECOGNITION RED TIDE ORGANISMS

作　　者：余肖翰[1] 谢杰镇[1] 郑少平[2] 

机构地区：[1]国家海洋局厦门海洋预报台,福建厦门361000 [2]厦门大学信息科学与技术学院,福建厦门361000

出　　处：《海洋湖沼通报》2011年第4期11-15,共5页Transactions of Oceanology and Limnology

摘　　要：本文研究了赤潮优势种识别技术,并实际运用于厦门大学流式细胞技术的赤潮实时监控系统FCAM。使用傅里叶形状特征描述子分析结合SVM算法作分类计算,通过对2009年4～10月厦门海域6种最常见的赤潮优势种的3000个样本为专门研究,并结合另外2种描述子算法将实验结果的识别精度提高到95.8%,具备较好的代表性。The recognition technology for the preponderant algae in red tide was studied and was applied to FCAM of Xiamen University real time monitoring system that using flow cytometry technology.Analysis for Fourier shape and feature descriptors and SVM was used as classifier,the experiment for classifying recognition precision that uniting with other two algorithms could be achieved to 95.8% about specially research of 6 classes 3,000 samples of Xiamen sea area＇s red tide algae.

关 键 词：形状特征描述子 傅里叶描述子 FCAM SVM-KNN 

分 类 号：TP3-05[自动化与计算机技术—计算机科学与技术]