基于MODIS植被指数时间谱的太湖2001年—2013年蓝藻爆发监测

发布时间：2018-05-06 05:22

本文选题：MODIS + 时谱　；参考：《光谱学与光谱分析》2016年05期

【摘要】：藻类水华爆发已成为影响内陆水体生态环境的重要因素。遥感能够提供实时的大范围观测,在水华监测中起到越来越重要的作用。遥感植被指数已广泛应用于藻类水华监测中,通过对研究区植被指数图像进行阈值分割,能够反映不同子区域内的藻类爆发程度;然而阈值分割法的结果只能反映某一时间点(图像获取时)的藻类爆发状况,无法表征长时间内藻类的变化。相比于单个时间点的植被指数,植被指数时间谱(时谱)包含藻类的物候信息,能够更加全面准确地反映藻类的长时间变化。目前,植被指数时间谱还尚未应用到水华相关研究中。选取2001年—2013年太湖区域的MODIS NDVI数据,构建年度NDVI时谱数据,利用(support vector machine,SVM)方法对每年的太湖蓝藻水华爆发强度进行分类,将太湖重度、中度和轻度蓝藻水华爆发的区域以及水生植物的区域提取出来,得到其空间分布和面积;并从2007年的时谱数据中抽取了8个时间点的NDVI图像,利用传统阈值分割法提取太湖重度、中度和轻度蓝藻水华爆发的区域,将结果与2007年时谱数据分类的结果进行对比。结果表明:所提出的方法能够更加全面准确地对太湖蓝藻爆发强度进行分类,通过NDVI时谱曲线提供的丰富物候信息可准确区分蓝藻与水生植被区域。本研究有望为准确掌握和预测藻类水华的爆发趋势及强度提供有效手段。
[Abstract]:The eruption of algae Shui Hua has become an important factor affecting the ecological environment of inland water bodies. Remote sensing can provide real-time wide-range observation and play a more and more important role in Shui Hua monitoring. Remote sensing vegetation index has been widely used in the monitoring of algae Shui Hua. Through threshold segmentation of vegetation index image in the study area, it can reflect the degree of algae eruption in different sub-regions. However, the results of threshold segmentation method can only reflect the algae explosion at a certain time point (image acquisition), and can not represent the changes of algae over a long period of time. Compared with the vegetation index of a single time point, the time spectrum of vegetation index (time spectrum) contains phenological information of algae, which can reflect the long-term variation of algae more comprehensively and accurately. At present, the time spectrum of vegetation index has not been applied to Shui Hua. The MODIS NDVI data of Taihu Lake region from 2001 to 2013 were selected to construct the annual NDVI time spectrum data, and the annual Shui Hua burst intensity of cyanobacteria was classified by using the support vector machine (SVM) method. The area of medium and mild cyanobacteria Shui Hua outbreak and the area of aquatic plants were extracted, and their spatial distribution and area were obtained. The NDVI images of eight time points were extracted from the time spectrum data of 2007. The regions of severe, moderate and mild cyanobacteria outbreaks in Taihu Lake were extracted by traditional threshold segmentation method, and the results were compared with the results of spectral data classification in 2007. The results show that the proposed method can more comprehensively and accurately classify the burst intensity of cyanobacteria in Taihu Lake, and the abundant phenological information provided by the NDVI time-spectrum curve can accurately distinguish the cyanobacteria from the aquatic vegetation area. This study is expected to provide an effective means to accurately grasp and predict the trend and intensity of algae Shui Hua outbreak.
【作者单位】：中国科学院遥感与数字地球研究所遥感科学国家重点实验室;中国科学院大学;
【基金】：国家自然科学基金项目(41201348,41371359) 高分水利遥感应用示范系统项目(08-Y30B07-9001-13/15-01)资助
【分类号】：X524;X87

【相似文献】