Sharp, Clancy (2019) Sugarcane Yield Estimation by UAV Photogrammetry Survey. [USQ Project]
|
Text (Project)
Sharp_C_Campbell_Redacted.pdf Download (6MB) | Preview |
Abstract
Sugarcane is a giant tropical grass in the botanical genus, Saccharum, the stalks of which are the world’s primary source of sugar (sucrose). After wheat, sugarcane is the second largest export crop in Australia with a total annual revenue of about $2.5 billion AUD.
There is a need for accurate and efficient yield estimation models for sugarcane crops, primarily because most of the cane is forward sold in the months leading up to harvest, and for logistical reasons including equipment allocation and harvest scheduling. Existing methods rely on hyperspectral satellite imagery and grower’s estimates, both of which have some limitations.
Unmanned aerial vehicles (UAVs), mounted with a visual spectrum (red-green-blue, i.e. RGB) camera may present an efficient and cost-effective method for capturing spatial and spectral data about sugarcane crops and, if processed and analysed properly, this data could be used to estimate the quantity of usable cane stalks in a canefield. Such a technique would be valuable for the sugarcane industry.
In this research project, the existing literature relating to crop height determination by UAV photogrammetry survey, visible-band spectral analysis of vegetation, and sugarcane yield estimation has been reviewed. A methodology was developed and a field study carried out to survey sugarcane crops using a consumer-grade UAV at approximately monthly intervals for three months leading up to harvest, to process the data into 3D digital models using photogrammetry software, to analyse the spatial and spectral properties of the data to find correlations with empirical yield data as recorded during a monitoring survey of the harvest, and to develop yield prediction models using linear regression and multiple linear regression techniques.
The results demonstrate that UAV-based photogrammetry is a suitable method to create digital models of the crop’s surface, and that the height of this surface model correlates strongly with empirical yield at all survey epochs. Such a technique is useful for assessing crop variability within fields. Unfortunately, however, mature cane is vulnerable to damage by wind and rain, which can affect its height and subsequently thwart observations about growth rate and yield predictions that are based on height. Visible-band vegetation indices exhibited low or erratic correlations with yield and were subject to influence from many factors including changing ambient light conditions and yellowing of the cane due to frost, thus rendering them an unreliable predictor of yield.
The conclusions of this project indicate promising potential for UAV photogrammetry survey in the sugarcane industry, with recommendations for future research to improve the yield prediction models by input of additional independent variables to overcome the obstacles discovered in this project.
|
Statistics for this ePrint Item |
| Item Type: | USQ Project |
|---|---|
| Item Status: | Live Archive |
| Faculty/School / Institute/Centre: | Historic - Faculty of Health, Engineering and Sciences - School of Civil Engineering and Surveying (1 Jul 2013 - 31 Dec 2021) |
| Supervisors: | Campbell, Glenn |
| Qualification: | Bachelor of Spatial Science (Honours) (Surveying) |
| Date Deposited: | 10 Aug 2021 06:03 |
| Last Modified: | 27 Jun 2023 04:04 |
| Uncontrolled Keywords: | Sugarcane, yield, estimation, Unmanned aerial vehicles (UAVs), UAV photography |
| URI: | https://sear.unisq.edu.au/id/eprint/43093 |
Actions (login required)
 |
Archive Repository Staff Only |
