论文部分内容阅读
基于遥感影像的农业大棚检测,能够快速获取大棚的空间分布情况与面积信息,对于农作物监测、农业规划等具有重要的意义。现有大棚遥感检测算法大多依赖于高分辨率遥感影像或无人机航空影像,存在成本高、算法设计复杂等不足。针对此问题,文章基于Landsat影像提出了一种快速大棚检测算法。首先,根据冬季大棚内表面产生冷凝露水这一常见自然现象,提出了一种增强型水体指数;然后结合归一化植被指数与可见光光谱特征,来更好地描述大棚及其他地物的特征。在此基础上,设计一种简单高效的决策树分类器识别大棚。文章以广东江门大鳌镇为例,对不同年份的Landsat影像展开实验并与其他方法进行对比。结果表明,文章所提方法有效地识别出了大棚,同时具有效率高、成本低、鲁棒性强的优点。
The detection of the agricultural greenhouse based on remote sensing images can quickly obtain the spatial distribution and the area information of the greenhouse and is of great significance for crop monitoring and agricultural planning. Most of the existing greenhouse remote sensing detection algorithms rely on high-resolution remote sensing images or UAV aerial images, which has the disadvantages of high cost and complex algorithm design. To solve this problem, this paper presents a fast greenhouse detection algorithm based on Landsat images. First, an enhanced water index is proposed based on the common natural phenomenon of condensation dew on the inner surface of winter greenhouse. Then, the characteristics of greenhouses and other features are better described by combining normalized vegetation index and visible light spectral characteristics. On this basis, we design a simple and efficient decision tree classifier to identify the greenhouse. The article takes the town of Da’ao, Jiangmen, Guangdong as an example to test Landsat images of different years and compare them with other methods. The results show that the method proposed in this paper can effectively identify the greenhouse and has the advantages of high efficiency, low cost and strong robustness.