.png){kind=icon}
返回
Automatic tree species recognition with quantitative structure models
.png){kind=icon}
PRE
AI
DOI：10.1016/j.rse.2016.12.002.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
We present three robust methods to accurately and automatically recognize tree species from terrestrial laser scanner data. The recognition is based on the use of quantitative structure tree models, which are hierarchical geometric primitive models accurately approximating the branching structure, geometry, and volume of the trees. Fifteen robust tree features are presented and tested with all different combinations for tree species classification. The classification methods presented are k-nearest neighbours, multinomial regression, and support vector machine based approaches. Three mainly single-species forest plots of Silver birch, Scots pine and Norway spruce, and two mixed-species forest plots located in Finland and a total number of trees over 1200 were used for demonstration. The results show that by using single species forest plots for training and testing, it is possible to find a feature combination between 5 and 15 features, that results in an average classification accuracy above 93% for all the methods. For the preliminary mixed-species forest plot testing, accuracy was lower but the classification approach presented potential to generalize to more diverse cases. Moreover, the results show that the post-processing of terrestrial laser scanning data of multi-hectare forest, from tree extraction and modelling to species classification, can be done automatically. (C) 2016 Elsevier Inc. All rights reserved.
Keyword:
Tree species recognition
Terrestrial laser scanning
Quantitative structure model
Tree reconstruction
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
11.4
论文数:
1.1W
被引数:
9.4W
机构
引用论文
.png){kind=icon}
分享收藏分享收藏Tree mapping using airborne, terrestrial and mobile laser scanning - A case study in a heterogeneous urban forest使用机载，地面和移动激光扫描进行树木制图-以异质城市森林为例
PREAI
分享收藏
查看更多内容