Abstract
The monitoring of the growth of crops in large-scale cultivation has relied on manual measurements of plant height
which is not only labour intensive but is also cumbersome. As an alternative, we proposed the utilisation of remote sensing techniques involving taking aerial images using Unmanned Aerial Vehicle (UAV) and it’s processing thereof. To process these images, a photogrammetric technique SfM (Structure from Motion) which involves estimating the three-dimensional structure from the multiple images obtained is utilised. However, to improve the accuracy of SfM it is imperative to properly position ground control points (GCPs) when taking images in the field. The placement of GCPs on the field is limited in large-scale cultivation where the entire field is covered by crops, especially inside the field as this would have negative impact on the uniformity of the crops and thus affecting the successive growth season. Our results indicated that by placing the GCPs at the edges of the field, i.e., at the corners and in between the corners, a total of 8 GCPs, we could precisely estimate the crop height, similarly to when more GCPs were placed inside the field. Since the UAV flight time is inversely proportional to the square of the flight altitude, a higher altitude would require less time, but result in lower spatial resolution. As a proof of concept, a flight altitude of 50 m was sufficient to obtain images of a 4-ha field with 8 GCPs on the edges of the field with high accuracy. These results indicated that UAVs combined with only GCPs on the edge of the field can be utilised to estimate with high precision the growth of crops in large-scale thus labour saving.
