摘要
采用基于单波段影像增强的ISODATA算法,以小麦和油菜为例进行实验.采用高分一号卫星影像作为数据源时,总体分类精度达到95.33%,比单独使用ISODATA算法提高约19%;采用Landsat 8影像作为数据源时,总体分类精度达到95.83%,比单独使用ISODATA算法提高约6%.分类结果的Kappa系数均在0.9 1.0之间,与原影像几乎完全匹配,有效解决农作物之间的混淆问题.该方法耗时短,精度高,可大范围推广.
Adopting a method of ISODATA algorithm which combining the enhancement of single band image and takes wheat and rape as examples. When using GF-1 satellite image as the data sources,comparing with the ISODATA algorithm,the overall classification accuracy increased by 19%,and reached to 95. 33%. At the same time,using the Landsat 8 image as the data source,the overall classification accuracy increased by 6%,reached to 95. 83%. The Kappa coefficients of classification results are all in 0. 9 1. 0,which almost completely matched with the original image,thus solved the problem of crop-confusion. In addition,the method was short time-consuming,high-accuracy and could be popularized in a large scale.
出处
《湖北大学学报（自然科学版）》
CAS
2017年第1期50-55,共6页
Journal of Hubei University：Natural Science
