About the project
University of Southern Queensland researchers are using drones to map farms to help form a more targeted, lower-cost and environmentally-friendly approach to eliminating weeds.
Large volumes of aerial imagery and drone data being collected by the USQ team are stored and managed on QCIF’s QRIScloud, a cloud computing service for Queensland researchers.
Weeds have a devastating economic impact on Australian agriculture. The federal Department of the Environment and Energy estimates weeds cost Australian farmers $1.5 billion a year in weed control activities and a further $2.5 billion a year in lost agricultural production.
Five researchers within USQ’s Centre for Agricultural Engineering (CAE) began working on the weed mapping project, an example of precision agriculture, in June 2017, with funding from the federal Department of Agriculture and Water Resources.
The CAE team is using machine vision approaches to identify weeds in fallow fields. Their aim is to develop a commercial system to reduce the use of herbicides when using existing weed-spraying technology. This will be achieved with drone-collected data and using software they have developed to automatically create weed spray maps based on this imagery. These maps will then be uploaded to a spraying device’s digital system so farmers can just target areas infested with weeds instead of the whole field.
Drone-use has resulted in a large amount of aerial imagery being generated for selected farms. This data is processed by a CAE team member to create orthomosaics of test fields. Orthomosaics involve an automated process of creating a mosaic of a number of images into one single, rescaled and georeferenced image that can be measured as a map is, and used as a digital surface model.
The CAE team is using QRIScloud to store the data collection. They are also using Nextcloud, QRIScloud’s research data management service, which provides users with a one-stop shop to access and manipulate data, using whatever device is preferred, be it laptop or desktop computer.
“We use QRIScloud to store all of this data so that it is all in one place, and Nextcloud, so it can quickly be shared with the rest of the team,” said Bruen Smith, a mechatronic engineer working on the project.
“The benefit of using QRIScloud is that we can organise and store a lot of data in one central location and we know that this data is safe. Also, most importantly, the data can be shared amongst the whole team and we know everyone is working with the same data set.”
QCIF’s USQ-based eResearch Analyst, Dr Francis Gacenga, assisted the research team with setting up QRIScloud and Nextcloud to securely store and synchronise the research data.
“Francis assisted us with setting up the repository as well as assisting us with some technical issues we were having at the beginning. Francis saved us a lot of time and confusion, enabling us to get on with our research,” said Mr Smith.
In collaboration with the University of Sydney, CAE has performed successful testing, literally in the field, of its weed spray mapping software in Narrabri, a rural NSW shire that has a major cotton-growing industry. The software successfully used map data to target particular areas of a field for weed-spraying. The research team is awaiting to collect data on how well this has done in controlling weeds.
The next steps will be to continue this testing using a variety of sprayers from different manufacturers, to investigate the use of other drone platforms to cover fields of more than 100 hectares, and to test the USQ-developed software’s ability to detect weeds on fields that have different crop stubble, i.e. the remnants after harvesting.
Weed spray mapping is just one of the projects USQ researchers are working on to help lead a technological revolution in Australian agriculture. Read more about USQ’s research in agriculture and agribusiness.
Bruen Smith collecting data in the field, using a real-time kinematic GPS device, with the help of project team member Livia Faria Defeo. (Image: Centre for Agricultural Engineering, USQ.)
Researcher: Bruen Smith
Centre for Agricultural Engineering
University of Southern Queensland
QRISdata: 1.34 TB. Frequent access storage.
One of the weed spray maps produced with USQ software, shown in Google Earth. (Image: Centre for Agricultural Engineering, USQ.)
Latest from Shannon Lindsay
- CQU data tool eases “mango madness”
- QRIScloud-powered app to increase profitability of Australia’s cattle industry
- QRIScloud supports globally significant Australian Threatened Species Index project
- National HPC access via QCIF paves way for researcher’s bright future
- QRIScloud delivers satellite data to Australian farmers, govt and the public