【24h】

AUTOMATIC DELINEATION OF FOREST STANDS FROM LIDAR DATA

机译:自动描绘森林的界限来自LIDAR数据

获取原文

摘要

Stands are the basic unit of forest management and information. The main disadvantage of stands is the labour-intensive updating work needed and the high variance in the quality of the results due to the amount of subjective judgment and manual work in creating the stands. In this study, a new approach for automatic stand delineation from a composite of LiDAR data - derived raster layers and colour-infrared aerial imagery - derived layer is introduced. The segmentation utilizes a new, iterative region-growing - based approach that forces the stands to be homogenous in timber type. The stand delineation quality was compared based on their ability to separate timber characteristics. The developed method was compared to two other stand delineation methods: (1) Automatically interpreted, with a segmentation algorithm in eCognition Pro 4.0 using LiDAR canopy height model. (2) Human-interpreted on aerial imagery (The traditional way in Scandinavian forestry). The testing was done on a 67-hectare forestland area in Juuka, Finland. 683 sample plots were laid on the property for control. This research shows it is possible to produce stand delineation automatically, utilizing LiDAR data, if timber characteristics are the only stand boundary criteria considered.
机译:代表是森林管理和信息的基本单位。展台的主要缺点是劳动密集型更新工作所需的工作,并且由于创建立场的主观判断和手动工作量,结果的高度方差。在这项研究中,引入了一种新的自动立式划分从激光雷达数据衍生的光栅层和颜色红外航空图像导出层的复合材料。分割利用新的迭代区域生长的方法,这些方法迫使站立在木材型中均匀。基于它们分离木材特性的能力,比较了立场划分质量。将开发的方法与另外两种立场描绘方法进行比较:(1)自动解释,使用激光轨道高度模型进行Ecognition Pro 4.0的分段算法。 (2)人类解释在空中图像(斯堪的纳维亚林业中的传统方式)。该测试是在芬兰Juuka的67公顷的林地区域完成。 683样品地块铺设在物业上进行控制。本研究表明,如果木材特征是所考虑的唯一站立边界标准,则可以自动生产立式描绘。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号