摘要
基于凉水国家级自然保护区2009年拍摄的1∶2 000航空像片和同期森林资源二类调查的固定样地数据,采用子像元分类方法分别提取出红松、落叶松和云冷杉的专题影像图。在此基础上,将栅格专题影像图转换为矢量图形,采用目视解译的方法提取上层针叶林的树冠信息。通过将针叶树冠形似为圆形提取出各树种的冠幅,用固定样地实测数据进行对比分析和精度评价,并建立航片上提取冠幅与实测冠幅之间的一元线性回归模型。结果表明:红松、落叶松和云冷杉冠幅的提取精度分别达到83.50%、84.35%和82.26%,其预测精度分别达到83.60%、81.46%和83.57%。
Based on 1 ： 2000 aerial photographs of Liangshui National Nature Reserve and permanent sample plots （PSPs） from forest resource inventory in 2009, the thematic images of Korean pine, larch, spruce and Amurfir were extracted respectively by the method of sub-pixel classification. In addition, crowns information of conifers in the over story were extracted by converting raster to vector and visual interpretation. On the assumption that the crown shape of conifer species was a circle, the crown diameters for each species were measured and the accuracy was validated by using field measured data from PSPs. A linear regression model between crown diameters extracted from the aerial photograph and the PSPs data was developed. Results show that the accuracy of crown diameter for Korean pine, larch, spruce and Amurfir is 83.50%, 84.35% and 82.26%, respectively, and the precision of prediction reaches to 83.6% , 81.46% and 83.57%, respectively.
出处
《森林工程》
2011年第6期16-22,共7页
Forest Engineering
基金
东北林业大学本科生创新实验项目(101022505)
关键词
子像元分类
栅格转矢量
冠幅提取
针叶树种
sub-pixel classification
raster to vector conversion
crown diameter extraction
conifer species
