一种快速区分易混农作物的遥感解译方法  

Remote Sensing Interpretation Approach for Fast Distinguishing Easilyconfused Crops

作　　者：许亮[1] 汪权方[1,2] 陈志杰[1,2] 王新生[1,2] 

机构地区：[1]湖北大学资源环境学院,湖北武汉430062 [2]农业部遥感应用中心武汉分中心,湖北武汉430062

出　　处：《地理空间信息》2017年第1期59-62,共4页Geospatial Information

基　　金：国家重点基础研究发展计划资助项目(2010CB950902);农业部遥感应用中心武汉分中心农业遥感监测与评价资助项目

摘　　要：在对冬小麦、油菜两种农作物进行遥感非监督分类时,二者很难区分,且分类精度不高。若先剔除原始遥感影像中的非冬小麦、油菜种植区,再使用ISODATA算法进行二次非监督分类,则可较容易地区分易混的冬小麦和油菜。结果表明,冬小麦二次分类精度比初始分类精度提高了20.6%,油菜二次分类精度比初始分类精度提高了19.4%,从而显著提高了农作物的遥感解译分类精度,大大减少了人工目视解译工作量。同时,该方法也为其他易混农作物的遥感解译工作提供了一种解决问题的思路。It is quite difficult to distinguish winter wheat and rape when taking unsupervised classification, and the classification accuracy is not high. Removing images of non-winter wheat and rape planting area from the original remote sensing images and using ISODATA to take the second unsupervised classification to planting area, could distinguish easily-confused winter wheat and rape rapidly. The results show that the accuracy of second classification of winter wheat is 20.6% higher than that of original classification, the accuracy of second classification of rape is 19.4% higher than that of original classification, which has dramatically reduced the workload of artificial visual interpretation. This approach could provide a reference for the remote sensing interpretation of similar easily-confused crops.

关 键 词：非监督分类 人工目视解译 二次分类 农业遥感 

分 类 号：P237[天文地球—摄影测量与遥感]