基于多特征的江苏海域赤潮遥感提取算法  

Jiangsu Sea of Red Tide Remote Sensing Based on Feature Extraction Algorithm

作　　者：彭模 刘寿东[1] 

机构地区：[1]南京信息工程大学应用气象学院

出　　处：《科技通报》2015年第3期25-28,53,共5页Bulletin of Science and Technology

基　　金：国家科技支撑计划项目(2013BAB04B03);江苏省科技支撑计划项目(BE2012774;BE2014729)

摘　　要：针对单一的赤潮特征提取算法在对赤潮遥感图像进行特征提取时还存在准确性不高等问题,本文提出了一种基于多特征的赤潮遥感提取算法,首先对加权平均方法进行改进,提出了自适应加权平均算法,对遥感图像的叶绿素浓度特征进行提取,然后采用偏差系数对水体固有光学模型进行参数优化,对遥感图像的光谱特征进行提取。实例仿真试验结果表明,本文提出的基于多特征的赤潮遥感提取算法相比较单一的光谱特征提取算法,具有较高的赤潮特征提取准确性和稳定性。Extraction algorithm for single characteristic red tide, not high accuracy in the presence of redtide remote sensing image feature extraction Shihai, this paper presents a red tide remote sensing basedon feature extraction algorithm, firstly, the weighted average method is proposed to improve the adaptivethe weighted average algorithm, the chlorophyll concentration characteristics of remote sensing imageswere extracted, and then using the coefficient of variation of water inherent optical model parameteroptimization, remote sensing images of spectral features were extracted. The simulation results show thatred tide remote sensing based on multi-feature extraction algorithm proposed in this paper compared tothe single spectral feature extraction algorithm, has a high tide feature extraction accuracy and stability.

关 键 词：赤潮特征提取 叶绿素浓度 水体固有光学模型 偏差系数 自适应加权平均 

分 类 号：P962[天文地球—自然地理学]