Figure 1.
Aerial view of the experimental field (a), showing the centers of the UAV aerial images in blue and the sampling points in black (see section 2.4), and in-field photograph of the study site (b), showing the maize rows and some patches of weed infestation.
Figure 2.
Unmanned quadrotor-type aerial vehicle flying over the crop field (a), and aerial image (color–infrared composition) obtained by the UAV at an altitude of 30 m (b), showing the maize rows, some weed patches and the Spectralon® panel.
Figure 3.
Flowchart of the OBIA procedure for classification of crop rows and weeds and generation of a weed infestation map.
Figure 4.
Partial view of the outputs of the OBIA procedure at each step: a) segmentation outputs at scales of 140 (in blue) and 10 (in black), used to calculate row orientation and define vegetation objects, respectively; b) classification of objects of vegetation and bare soil ; c) definition of the crop row structure (in black); d) classified image with crop, weeds and bare soil; e) grid framework of the inter-row area; f) weed coverage map showing three levels of infestation (low, moderate and high), crop rows and weed-free zones.
Figure 5.
On-ground photographs (1) and UAV images (2) of the 1x1-m frames used in the ground-truth sampling of three different categories of weed coverage: a) low, b) moderate, and c) high.
Figure 6.
Observed and estimated weed coverage (%) inside the sampling frames from on-ground photographs and UAV image analysis, respectively.