基于无人机影像的太原东山实验林场油松林冠幅提取及蓄积量反演  

Canopy Extraction and Volume Inversion of Pinus Tabulaeformis Forest in Taiyuan Dongshan Experimental Forest Farm Based on UAV Images

作　　者：范晓龙[1] 李晓迪 郭媛 张引[1] Fan Xiaolong;Li Xiaodi;Guo Yuan;Zhang Yin(Shanxi Forestry Vocational Technical College, Taiyuan 030009, China)

机构地区：[1]山西林业职业技术学院,山西太原030009

出　　处：《山西林业科技》2022年第2期25-27,共3页Shanxi Forestry Science and Technology

摘　　要：基于山西省太行山国有林管理局东山实验林场的无人机影像数据和油松样方的每木检尺数据,利用ENVI 5.3软件对无人机影像进行多尺度分割,确定最佳的冠幅分割参数。采用面向对象的分类方法提取油松影像冠幅,通过影像冠幅与实测胸径构建冠幅-胸径模型,并进行蓄积量反演。结果表明,当分割尺度设为35,合并尺度设为80时,冠幅边缘较为清晰,对影像的分割较为合理。胸径-冠幅估计模型以一元线性模型拟合度最高,决定系数R^(2)为0.7797,将提取参数代入模型反演蓄积量,可得到航空立木材积表。Based on the UAV image data of Dongshan experimental forest farm of National Forest Administration Bureau of Taihang Mountain and the data of each timber scaling of Pinus tabulaeformis sample plots,the UAV image was multi-scale segmented by ENVI 5.3 software to determine the best crown segmentation parameters.The object-oriented classification method was used to extract the image crown width of Pinus tabulaeformis,and the crown width-DBH model was constructed through the image crown width and the measured DBH to inverse the volume.The results showed that when the segmentation scale parameter was set to 35 and the merging scale was set to 80,the crown width edge was clearer and the image segmentation was more reasonable.The fitting degree of the DBH-crown width estimation model was the highest with the univariate linear model,and the coefficient of determination R2 was 0.7797.By substituting the extracted parameters into the model to inverse the volume,the air standing timber volume table could be obtained.

关 键 词：无人机影像 油松 冠幅 面向对象分类方法 蓄积量 

分 类 号：S771.5+4[农业科学—森林工程]