欢迎访问《中国农学通报》，
中国农学通报

中国农学通报 ›› 2013, Vol. 29 ›› Issue (1): 47-52.doi: 10.11924/j.issn.1000-6850.2012-3098

• 林学 园艺 园林 • 上一篇    下一篇

不同分类方法在竹林遥感信息识别中的应用

官凤英 范少辉 郜燕芳 肖复明 蔡华利   

  • 收稿日期:2012-09-11 修回日期:2012-10-08 出版日期:2013-01-05 发布日期:2013-01-05
  • 基金资助:
    国际竹藤中心科研专项“竹资源遥感专题信息提取技术研究”(1632010012)；“十二五”国家科技支撑计划项目“竹林资源监测与管理技术”(2012BAD23B04)；林业科技成果推广项目“竹资源遥感监测及信息化管理技术示范推广”[(2012)36号]；江西省财政林业重大专项“江西典型人工林生态系统碳循环特征及其调控技术研究”(2011511101)。

Different Classification Methods in the Application of Remote Sensing Information Identification of Bamboo

  • Received:2012-09-11 Revised:2012-10-08 Online:2013-01-05 Published:2013-01-05

PDF (PC)

778

摘要： 应用遥感技术开展竹资源监测与调查，能快速、高效地为竹资源的科学管理和高效利用提供基础数据支撑。利用福建省顺昌县TM影像，采用最大似然法、子像元分类和光谱特征分类3种方法进行竹林信息提取研究。结果表明：子像元分类总体精度最高为77.33%，基于光谱特征分类和最大似然法的总体精度分别为76.50%、76.17%；3种分类方法的Kappa系数分别为72.8%、71.8%、71.4%；与顺昌县2007年森林资源二类清查竹林面积进行比较，光谱特征分类法精度最高为95.68%，最大似然法和子像元分类法的精度分别为93.41%、92.97%。

关键词: 脂肪酸, 脂肪酸

Abstract: The application of remote sensing technology to monitor and investigate the bamboo resource can provide basic data support for the scientific management and efficient utilization of the bamboo resource quickly and effectively. The author used the TM images of Shunchang County Fujian Province, using the maximum likelihood method, the sub-pixel classification and the classification based on spectral characteristics 3 classification methods to extract the bamboo information. The result showed that: the sub-pixel classification was the most precise, and the overall accuracy up to 77.33% , followed by the classification based on spectral characteristics, maximum likelihood method, the accuracy respectively was 76.50% and 76.17% ; the Kappa coefficient respectively was 72.8% , 71.8% and 71.4% ; and compared the results of the 3 classification methods to the bamboo forest area of the second survey in 2007, the accuracy of the classification based on spectral characteristics was the highest 95.68%, followed by maximum likelihood method 93.41%, the sub-pixel classification 92.97%.