论文部分内容阅读
在覆盖玉米田的遥感图像中,玉米地块边缘区域存在大量的同物异谱现象,利用传统分割方法进行玉米地块分割时,会造成边缘出现许多非玉米小块区域,因而导致玉米种植面积统计错误。根据大面积玉米种植区域的形状分布特点,提出一种类矩形引导的玉米田种植区分割方法。首先采用最小核值相似区(smallest univale segment assimilating nucleus,SUSAN)边缘检测算子对融合后的高分一号(GF-1)卫星遥感图像进行边缘提取,然后根据闭合区域与外接类矩形的关系构建类矩形引导的相关函数,最后将类矩形阈值函数引入基于图的分割算法中实现特定形状的地块分割。将分割结果分别与基于图的分割方法、分水岭分割方法和人工解译样本进行实验比较,结果表明:本文方法能有效地分割出玉米田目标,减少了同物异谱带来的影响,分割结果更加符合玉米田实际分布特征和实际统计面积。
In the remote sensing images covering maize fields, there are a lot of same-matter phenomenon in the marginal area of corn plots. When the traditional segmentation method is used to divide the maize parcel, many non-corn small areas appear on the edge, which leads to the maize acreage Statistics wrong. According to the shape distribution characteristics of large-area corn planting area, a rectangle-oriented planting area division method was proposed. Firstly, edge detection is performed on the fused high score one (GF-1) satellite remote sensing image using the edge detection operator of the smallest univale segment assimilating nucleus (SUSAN). Based on the relationship between the closed region and the circumscribed rectangle Constructing the class rectangle-oriented correlation function, and finally introducing the class rectangle threshold function into the graph-based segmentation algorithm to achieve a specific shape of the parcel. The results are compared with those based on graph segmentation, watershed segmentation and artificial interpretation respectively. The results show that the proposed method can effectively separate the target of corn field and reduce the influence of the same spectrum. The segmentation result More in line with the actual distribution of corn fields and the actual statistical area.