基于机载激光雷达数据的老街基林场样地单木分割方法研究

Individual-tree Segmentation Method in the Sample Plot of Laojieji Forest Farm Based on Airborne LiDAR Data

机载激光雷达技术凭借着对森林冠层的穿透能力,在森林资源调查中显露出无可比拟的优势。针对林业雷达领域中的单木结构参数提取问题,本文基于局部最大值滤波算法和标记控制分水岭分割算法(Marked-Controlled Watered Segmentation,MCWS)提出了一种单木分割方法,并评价该方法在样地尺度上的单木分割效果,分析结果表明样地区域总体分割精度F为0.736;结合实测树高对反演树高进行相关性分析,拟合精度R^(2)为0.907,表明本文方法能够较准确地提取单木树冠信息。

Airborne LiDAR technology shows incomparable advantages in forest resource survey by virtue of its penetrating ability to forest canopy.Aiming at the problem of individual-tree structure parameter extraction in the field of forestry LiDAR,this paper propo-ses a individual-tree segmentation method,which is based on the local maximum filtering algorithm and the Marked-Controlled Wa-tershed Segmentation(MCWS)algorithm and evaluates the effect of individual-tree segmentation on the scale of sample plots,the a-nalysis results show that the overall segmentation accuracy F of the sample region is 0.736.Combining the measured tree height with the correlation analysis of the inverted tree height,the fitting accuracy R^(2) is 0.907,indicating that the method in this paper can extract the crown information more accurately.