Parameterization of the Individual Tree Detection Method Using Large Dataset from Ground Sample Plots and Airborne Laser Scanning for Stands Inventory in Coniferous Forest

Remote Sensing

Łukasz Kolendo¹  Marcin Kozniewski²  Szymon Chmur³  Bożydar Neroj³  Marek Ksepko³ 

[1] Faculty of Civil Engineering and Environmental Sciences, Bialystok University of Technology, Wiejska 45E, 15-351 Bialystok, Poland;Faculty of Computer Science, Bialystok University of Technology, ul. Wiejska 45A, 15-351 Bialystok, Poland;Forest Management and Geodesy Bureau, Sekocin Stary, ul. Lesnikow 21, 05-090 Raszyn, Poland;

关键词: precise forestry;    precise forest management;    LiDAR;    ALS data;    tree segmentation;    tree tops;   

DOI  :  10.3390/rs13142753

来源: DOAJ

【 摘 要 】

Highly accurate and extensive datasets are needed for the practical implementation of precision forestry as a method of forest ecosystem management. Proper processing of huge datasets involves the necessity of the appropriate selection of methods for their analysis and optimization. In this paper, we propose a concept for and implementation of a data preprocessing algorithm, and a method for the empirical verification of selected individual tree detection (ITD) algorithms, based on Airborne Laser Scanning (ALS) data. In our study, we used ALS data and very extensive dendrometric field measurements (including over 21,000 trees on 522 circular sample plots) in the economic and protective coniferous stands of north-eastern Poland. Our algorithm deals well with the overestimation problems of tree top detection. Furthermore, we analyzed segmentation parameters for the two currently dominant ITD methods: Watershed (WS) and Local Maximum Filter with Growing Region (LMF+GR). We optimized them with respect to minimizing the Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). Additionally, our results show the crucial importance of the quality of empirical data for the correct evaluation of the accuracy of ITD algorithms.

【 授权许可】

Unknown