基于MODIS数据黄海绿潮覆盖面积精细化提取

发布时间：2018-03-19 10:05

本文选题：黄海　切入点：绿潮　出处：《激光生物学报》2017年01期 　论文类型：期刊论文

【摘要】：针对MODIS数据绿潮提取存在大量混合像元导致绿潮覆盖面积偏大这一问题,基于3 m分辨率的机载SAR高分影像,结合归一化植被指数(NDVI),对MODIS研究区中大于NDVI阈值的像元进行混合像元分解得到绿潮的"覆盖面积";获取的准同步3 m机载SAR提取的绿潮面积为"真实值",然后建立二者之间的关系模型,并选取不同的样本区域对该模型进行了验证。实验结果表明:NDVI等传统算法所提取的绿潮覆盖面积约为"真实值"的2.68倍;基于混合像元分解的方法所提取的绿潮面积较"真实值"偏小,约为"真值"的0.56倍;与传统的NDVI等多波段比值法相比,该精细化模型方法提取的绿潮覆盖面积更接近于"真实值",与"真实值"误差仅为6.7%。
[Abstract]:In order to solve the problem that a large number of mixed pixels lead to a large green tide coverage area in MODIS data extraction, the airborne SAR high score image based on 3 m resolution is proposed. Combined with normalized vegetation index (NDVI), the mixed pixel decomposition of pixels larger than NDVI threshold in MODIS study area is used to obtain the "coverage area" of green tide, and the area of green tide extracted by quasi-synchronous 3 m airborne SAR is "real value", and then constructed. Establish a model of the relationship between the two, The model is verified by selecting different sample areas. The experimental results show that the green cover area extracted by the traditional algorithms such as: NDVI is about 2.68 times of the "true value". The area of green tide extracted by the method based on mixed pixel decomposition is smaller than that of "true value", which is about 0.56 times that of "true value", which is compared with the traditional multi-band ratio method such as NDVI. The green cover area extracted by the refined model method is closer to the "real value", and the error between the "true value" and the "true value" is only 6.7.
【作者单位】：中国石油大学地球科学与技术学院;
【基金】：国家自然科学基金“主被动光学遥感探测水下悬浮绿潮”(41476159);国家自然科学基金“海洋灾害大数据分析的系统模型研究及应用”(41476101)
【分类号】：X55;X87

【相似文献】