Detection of necrotic foliage in a young Eucalyptus pellita plantation using unmanned aerial vehicle RGB photography – a demonstration of concept

Figures & data

Table 1. Alignment parameters set for processing the photographs acquired by the DJI Phantom 4 Pro for the survey over a planted stand of young Eucalyptus pellita

Processing parameters	Values
Alignment
Accuracy	High
Generic preselection	Yes
Reference preselection	No
Key point limit	40 000
Tie point limit	4000
Adaptive camera model fitting	Yes
Matching time	11 min 52 s
Alignment time	3 min 11 s
Optimisation
Parameters	F, b1, b2, cx, cy, k1-k4, p1, p2
Adaptive camera model fitting	No
Optimisation time	8 s
Dense point cloud
Points	16 653 614
Point colours	3 bands (RGB), unit 8
Reconstruction
Quality	High
Depth filtering	Mild
Depth maps generation time	2 h 34 min
Dense cloud generation time	3 h 51 min

RGB, Red Green Blue.

Figure 1. Individual tree heights annotated onto an orthophoto derived using a digital surface model (DSM). High points in the DSM are closely aligned with the image of each tree top

Figure 2. Illustration of settings used in the (a) Agisoft photoscan set raster transform function and (b) raster transform palette settings used for VARI VI VARI, visible atmospheric resistant index

Figure 3. (a) The RGB point cloud of a small stand of trees from within the study site; (b) shows the corresponding normalised and Z clipped VARI-transformed point cloud

RGB, Red Green Blue; VARI, Visible Atmospheric Resistant Index

Figure 4. Comparison of a partially diseased Eucalyptus pellita crown as presented in the RGB point cloud dataset with the enhanced point cloud dataset transformed through application of the VARI algorithm to the individual points

RGB, Red Green Blue; VARI, Visible Atmospheric Resistant Index

Figure 5. Nadir view of the VARI point cloud that has been height normalised and ‘Z clipped’ to remove the influence of the understorey and soil influences. The red dots represent necrotic trees

VARI, Visible Atmospheric Resistant Index

Table 2. Confusion matrix and accuracy results for visually detecting healthy, partially affected, dead and missing Eucalyptus pellita crowns affected by the bacterial wilt Ralstonia sp

RGB orthophoto
		Healthy	Partially necrotic	Completely necrotic	Missing	Total
VARI-transformed orthophoto	Healthy	1131	7	0	0	1138
	Partially necrotic	2	28	0	0	30
	Completely necrotic	0	2	40	0	42
	Missing	0	2	0	38	40
	Total	1133	39	40	38	1250

RGB, Red Green Blue; VARI, Visible Atmospheric Resistant Index.