基于云覆盖分类的太湖蓝藻MODIS图像分析

Cyanobacteria Information Extraction in Taihu Lake Based on Cloud Analysis With MODIS Imageries

太湖蓝藻已经成为一个社会和政府共同关注的环境问题,如何及时、准确地获取蓝藻发生、发展状况信息,是目前水环境监测的难点.本文以MODIS遥感图像为监测信息源,提出一种新的彩色图像分析方法快速提取太湖蓝藻信息.首先使用模板匹配算法快速提取MODIS遥感图像的太湖水域信息,并针对遥感图像获取过程中的云层覆盖干扰,提出将图像分为无云覆盖和有薄云覆盖两类分别处理的方法,提高了不同天气情况下蓝藻信息提取的有效性.最后通过对一年内多幅太湖MODIS遥感图像的蓝藻信息提取,定性分析了太湖蓝藻的时空分布特点,为研究太湖蓝藻发生、发展规律提供技术支撑.

Cyanobacteria bloom-forming in Taihu Lake has already become an environmental problem which is concerned with the whole society and the government. It is difficult for the department of water-environment monitoring to monitor the bloom on the lake timely and accurately. By using template matching process, the territorial waters information of Taihu Lake is extracted firstly. To tackle the problem of Cyanobacteria under clouds cover, the problem is divided into two, one is under the clouds cover, the other is under no clouds. Experimental results on hundreds of MODIS images show that the proposed method is effective with satisfactory accuracy. Furthermore, spatio-temporal analysis of Cya- nobacteria during the year 2006 reveals the underlying knowledge of its year pattern, which can provide decision support information for the administrative departments.