基于遥感影像光谱分析的蓝藻水华识别方法

Recognition of Cyanobacteria Bloom Based on Spectral Analysis of Remote Sensing Imagery

摘要:

利用Landsat7 ETM+遥感影像数据,以淀山湖为例,在分析蓝藻和其他典型地物影像光谱曲线及其特征的基础上,构建归一化蓝藻指数(NDI_CB),有效地从浑浊水体中提取蓝藻信息.通过k均值非监督分类结果可以发现,构建的归一化蓝藻指数较传统的归一化差值植被指数(NDVI)和比值植被指数(RVI)更加适用于提取低密度蓝藻空间分布信息.在此基础上,基于遥感影像光谱特征和归一化蓝藻指数,采用了支持向量机的分类识别模型,最终得到淀山湖区域蓝藻的空间分布范围与面积,通过发现在某一特定时间蓝藻分布的规律,为蓝藻预警和治理的生态学分析提供了及时、有效和客观的依据.

Abstract:

Based on the analysis of spectral curve and features of cyanobacteria bloom and other typical ground object,the normalized difference cyanobacteria bloom index(NDI_CB)was constructed to distinguish between cyanobacteria bloom and turbid water with the Landsat7 ETM+ image in Lake Dianshan.In this study two other different vegetation indexes,normalized difference vegetation index(NDVI)and ratio vegetation index(RVI),together with NDI_CB,were applied to extracting the cyanobacteria bloom information from the same image via unsupervised classification method(kmeans).The results show that NDI_CB is the best one for lowdensity cyanobacteria bloom extraction.In order to recognize the cyanobacteria bloom better,support vector machine(SVM)classification method was used to classify the image based on spectral features and NDI_CB,and to obtain the spatial distribution and the area of cyanobacteria bloom in Lake Dianshan.Through studying the laws of the cyanobacteria bloom distribution at a particular time,a sound,efficient and objective basis has been achieved for the ecological analysis of the prevention and the treatment of cyanobacteria bloom.