南京林业大学学报(自然科学版) ›› 2000, Vol. 24 ›› Issue (03): 49-52.doi: 10.3969/j.jssn.1000-2006.2000.03.016

• 研究论文 • 上一篇    下一篇

数字化CIR图像专题特征的提取及成图

徐世永;林杰;陈涛   

  1. 南京林业大学;南京210037;南京林业大学;南京210037;洛阳林业学校
  • 出版日期:2000-06-18 发布日期:2000-06-18

A Study on Forest Type Mapping with Digital Colour Infrared Photograph

Xu Shiyong Lin Jie ?Cheng Tao(Nanjing Forestry University\ Nanjing 210037)   

  1. Luoyang Forestry School
  • Online:2000-06-18 Published:2000-06-18

摘要: <正>通过对南京紫金山地区 1 996年 1∶2万的彩色红外航片数字影像进行多种方法的图像增强处理表明 ,采用红绿两个单波段进行归一化差值 (R -G/R +G)处理和红绿蓝 3个波段进行KL变换 ,能够最大限度地提取森林植被信息。把非监督分类和监督分类有机地结合起来 ,可实现森林类型的自动分类 (极大似然法的分类精度为 87.4 %) ,并输出森林类型分布图

Abstract: Forest thematic mapping was of high cost work. But now we can promote the efficiency greatly with the combination of remote sensing and computer technology. The 1∶20?000 colour infrared photograph in Zijing mountain area of 1996 was enhanced with many methods. It was shown that forest information could be most extracted using NDVI and KL transformation. Furthermore, the forest types were classified automatically by combining non-supervised and supervised methods, and the classification accuracy reached 87.4%.Finally,the forest type map was output. So we can make a map of new forest types rapidly, gain some help to manage this landscape area well using the modern mapping method.

中图分类号: