面向对象方法支持下的农业耕作变化检测—以黑龙江肇州县为例

Agricultural Change Detection Supported by Object-Oriented Method—Zhaozhou County in Heilongjiang as an Example

摘要: 随着遥感技术的发展，遥感在农业方面的应用就更加广泛深入。本文选取Landsat系列卫星1990~2015五个年代的影像数据，采用面向对象的方法对黑龙江省肇州县进行农业耕作变化检测。影像选取7月至9月农作物长势较好的时间，肇州县农田规划和分布十分有规律，遥感影像纹理清晰，适用于面向对象的方法进行检测。本文通过对对象的目视识别来选取样本得到分类结果，再利用分类后比较法对农业耕作进行变化检测。结果表明，肇州县在1990~2000年耕地的面积呈增长趋势；2000~2005年、2005~2010年、2010~2015年、1990~2015年耕地的面积呈减少趋势。

Abstract: With the development of remote sensing technology, remote sensing application in agriculture is even more extensive. This paper selects the image data of Lands series satellite during 1990~2015, and adopts the object-oriented approach to detect the change on agricultural cultivation in Zhaozhou County, Heilongjiang Province. The image is selected from July to September, when crops are growing well. Planning and distribution of Zhaozhou County farmland are very regular. Remote sensing image texture is clear, which is suitable for testing in the object-oriented approach. Classification result is obtained by visual identification of objects to select samples, after which the changes of farming are detected using comparison method. The results show that the area of cultivated land in Zhangzhou County increased from 1990 to 2000, and decreased in 2000~2005, 2005~2010, 2010~2015 and 1990~2015.