利用3×3卷积的遥感图象分类系统
The Classification System for Remote Sensing Image Using 3×3 Convolution
-
摘要: 本文对图象纹理分析提出了一些新的3×3卷积模板，构成了实用的遥感图象分类系统；基于微机实验，在七种地物类型的245个样本窗口中鉴别，获得了高达99.18%的识别正确率．Abstract: In this paper,texture analysis for remotely sensed image is developed by using our new 3×3 convolution masks;A prectical tree-type classifier is formed;Image recognition experiment based on micro-computer,245 sample windows (7 class of terrains) is performed,and a correct rate,99.18%,is acquired.
下载:
