基于改进的NDVI密度分割方法的冬小麦面积信息提取

发布时间：2018-04-01 07:27

本文选题：NDVI　切入点：高分一号　出处：《南水北调与水利科技》2017年03期

【摘要】：基于高分辨率遥感影像提取的种植结构信息,能够比传统的统计数据更加直观地表达农作物的空间分布特征,这些数据信息是水资源管理部门进行水资源管理的重要数据参考。为解决GF-1 WFV传感器影像中混合像元对小麦信息提取结果的影响,引入高分辨率GF-1PMS传感器影像,在两种影像相同位置建立样本研究区,利用PMS影像的分辨率优势为WFV影像中小麦混合像元训练样本提供真实小麦面积权重,得到WFV影像小麦混合像元NDVI与小麦面积权重的比例关系,再运用区间归一化的方法解决同一NDVI值对应不同小麦面积权重的问题,进而得到混合像元中小麦的真实面积信息,最终提取了冀州市的冬小麦信息。经验证,该方法能够在实地样本不足的条件下,较准确地提取冬小麦面积信息。
[Abstract]:The planting structure information extracted from high resolution remote sensing images can express the spatial distribution characteristics of crops more intuitively than the traditional statistical data. These data information is an important data reference for water resources management department.In order to solve the influence of mixed pixels in GF-1 WFV sensor image on wheat information extraction, high resolution GF-1PMS sensor image was introduced, and the sample study area was established in the same position of two kinds of images.Using the resolution advantage of PMS image to provide real wheat area weight for wheat mixed pixel training sample in WFV image, the relation between wheat mixed pixel NDVI and wheat area weight in WFV image is obtained.Then the method of interval normalization is used to solve the problem that the same NDVI value corresponds to different weight of wheat area, and then the real area information of wheat in mixed pixel is obtained, and finally the information of winter wheat in Jizhou City is extracted.It is proved that this method can accurately extract the area information of winter wheat under the condition of insufficient field samples.
【作者单位】：兰州交通大学测绘与地理信息学院;中国水利水电科学研究院;
【基金】：国家科技重大专项(08Y30B07-9001-13/15) 国家科技支撑计划项目(2013BAB05B01)~~
【分类号】：S127;S512.11

【相似文献】