研究目的
To evaluate the use of unmanned aerial systems (UASs) for 3D forest structure mapping inside forest stands using photogrammetry techniques, specifically assessing the accuracy of diameter at breast height (DBH) estimation.
研究成果
The study demonstrated that UASs can successfully map forest structure interiors, with the Phantom 4 Pro providing more accurate DBH estimates than the Mavic Pro. The least squares ellipse method yielded the lowest errors, and accuracy improved with increased image coverage. This approach offers a low-cost alternative to traditional methods and does not require ground control points, simplifying data acquisition.
研究不足
The method requires open space for flight, limiting it to even-aged stands with sufficient distance between trees (minimum 2m), higher crown bases, and no high understory or thin branches. It may not be applicable in all forest types, and manual piloting was necessary due to irregular tree distribution.
1：Experimental Design and Method Selection:
The study utilized close-range photogrammetry with UASs to create 3D point clouds of forest stands. Flights were manually piloted in a double zig-zag pattern at 8 m height above ground.
2：Sample Selection and Data Sources:
Two 50m x 50m research plots in mature Norway spruce and European beech stands in the Czech Republic were used. Field measurements included tree positions and DBH using laser rangefinders, total stations, and hypsometers.
3：List of Experimental Equipment and Materials:
UASs (DJI Phantom 4 Pro and DJI Mavic Pro), cameras integrated into UASs, Agisoft PhotoScan software, MATLAB R2017b, TruPulse 360 laser rangefinder, TOPCON GTS-210 total station, VL5 Vertex Laser hypsometer.
4：Experimental Procedures and Operational Workflow:
Images were acquired with timed shot intervals, processed using PhotoScan for point cloud generation, and DBH was estimated using four fitting methods (bounding circle, convex hull, least squares circle, least squares ellipse). Accuracy was assessed by comparing with field-measured DBH.
5：Data Analysis Methods:
Statistical analysis included bias and RMSE calculations, N-way ANOVA to test effects of method, sensor, and species, and linear regression to analyze error dependencies on image coverage.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
