摘要
提出了一种综合利用小波变换高低通滤波数据的海洋赤潮识别方法。基于低通滤波数据，利用基于有限混合密度理论期望最大（EM）算法作为最大似然分类（佃屺）参数估计的方法（EM-MLC）来进行赤潮、非赤潮和过渡水体的分类识别，并可进一步识别出不同优势种藻类引发的赤潮区域；利用高通滤波数据，可以分析赤潮爆发中非优势种藻类的信息，这就为引发赤潮的藻类种类的判断奠定了基础。通过实验验证了本方法可以有效地进行赤潮识别。同时，根据检测出的过渡水体区域信息，可以进行赤潮爆发前的预测。
This paper presents an ocean red tide recognition method based on the data from the high and low pass filter of the wavelet decomposition. Based on the lowpass data,the red tide, the non-red-tide and the transitional seawater areas can be recognised through the method that uses the Expectation Maximization （EM） algorithm to estimate the parameters of the conventional Maximum Likelihood Classification （MLC） EM-MLC; the red tide areas that are resulted by the different dominant species respectively can be recognised ulteriorly. The information of the non-dominant species in the course of the red tide can be analyzed using the highpass data, which establishes a foundation to estimate the sorts of the algae. The comparative experiments prove that the method can recognise the red tide effectively. At the same time, the forecast of the red tide can be done according to the recognized area of the transitional seawater.
出处
《高技术通讯》
CAS
CSCD
北大核心
2006年第6期652-656,共5页
Chinese High Technology Letters
基金
863计划（2001AA-636030）、青岛科技大学博士基金资助项目.
关键词
EM-MLC
小波分解
海洋赤潮
高光谱图像
赤潮识别
EM-MLC, wavelet decomposition, ocean red tide, hyper-spectral image, red tide recognition
