The Use of a Hand-Held Camera for Individual Tree 3D Mapping in Forest Sample Plots

Xinlian Liang,Juha Hyyppä,Jingbin Liu,Harri Kaartinen,Anttoni Jaakkola,Eija Honkavaara,Yunsheng Wang

doi:10.3390/rs6076587

Xinlian Liang, Juha Hyyppä + Show 5 more

Open Access

https://doi.org/10.3390/rs6076587

Copy DOI

Abstract

This paper evaluated the feasibility of a terrestrial point cloud generated utilizing an uncalibrated hand-held consumer camera at a plot level and measuring the plot at an individual-tree level. Individual tree stems in the plot were detected and modeled from the image-based point cloud, and the diameter-at-breast-height (DBH) of each tree was estimated. The detected-results were compared with field measurements and with those derived from the single-scan terrestrial laser scanning (TLS) data. The experiment showed that the mapping accuracy was 88% and the root mean squared error of DBH estimates of individual trees was 2.39 cm, which is acceptable for practical applications and was similar to the results achieved using TLS. The main advantages of the image-based point cloud data lie in the low cost of the equipment required for the data collection, the simple and fast field measurements and the automated data processing, which may be interesting and important for certain applications, such as field inventories by landowners who do not have supports from external experts. The disadvantages of the image-based point cloud data include the limited capability of mapping small trees and complex forest stands.

Highlights

IntroductionReference data collected from sample plots are fundamental for forest inventory and ecology
Reference data collected from sample plots are fundamental for forest inventory and ecology.Plot- and individual-tree-level measurements are made in sample plots
The result of this study showed that DBHs and locations of trees within the forest plot can be estimated from image-based point clouds at an acceptable level for practical field inventories

Summary

Introduction

Reference data collected from sample plots are fundamental for forest inventory and ecology. Plot- and individual-tree-level measurements are made in sample plots. Plot-level measurement takes a sample plot as a minimum unit and describes the plot using average values, whereas individual-tree-level measurement collects the attributes of each tree. The latter has the advantage of providing more accurate and comprehensive information regarding a sample plot. If tree locations were measured, individual trees in the sample plots can be connected to high-resolution remote sensing (RS) data, which can be utilized to teach RS data to carry out precision forest inventories over large areas.

Methods

Results

Discussion

Conclusion