This International Open Access week the Library is highlighting the University’s ongoing efforts in supporting open scholarship and knowledge sharing.
Open data and the FAIR principles (Findable, Accessible, Interoperable and Reusable) bring significant benefits to researchers, practitioners, industry and the public. When research data is accessible, it democratises information, reduces costs, minimises duplication, enhances research transparency and opens doors to innovation.
Today, we’re highlighting Griffith University’s research on sugar cane diseases, where researchers have made their data openly available to benefit the community.
Why sugar cane matters
James De Mers Pixabay
Sugar cane is a vital industry in Queensland and Australia, supporting jobs, businesses and, of course, satisfying our sweet tooth! However, diseases like sugarcane smut pose serious threats to this industry. If left unchecked, sugarcane smut can lead to production losses between 30% and 100%. Early detection is crucial since this disease is difficult to spot in its initial stages.
To address this challenge, Professor Jun Zhou, PhD candidate Dong Bau and their team developed a method using hyperspectral imaging to automatically detect the early symptoms of sugarcane smut.
Why share the data?
We asked Professor Zhou about the significance of sharing this research data:
‘To our knowledge, this is the first open access, large-scale hyperspectral image dataset for sugarcane disease detection. Building this dataset required access to specialised equipment, knowledge in hyperspectral imaging machine learning, and facilities for growing and monitoring sugarcane diseases’…
Meeting these requirements was a significant effort from Griffith and Sugar Research Australia (SRA). The goals for sharing the data are to:
- Promote hyperspectral imaging research in agriculture.
- Enable the development of new technologies for detecting sugarcane diseases.
- Enhance Griffith University’s reputation in AI and agricultural research.
- Attract collaboration opportunities and extend our research impact.
Since its release in 2023, the dataset has attracted global interest, with access requests from researchers in countries including China, India, Brazil and Australia. Many are using the data to train deep learning models or to explore hyperspectral imaging for plant disease detection. One project even involved creating a free app for farmers.
Publishing and sharing your data
Griffith Research Online (GRO) offers a streamlined way for researchers to publish their data. GRO Research Data Collection is supported by linked storage with Griffith’s Research Storage Services and other approved locations. Datasets such as the Hyperspectral Image datasets can also receive a Digital Object Identifier (DOI) to enhance discoverability and citation tracking.
If you’re interested in sharing your research data, explore the Library’s services tools and training resources You can also reach out to the Library for guidance on using GRO or other trusted research data repositories.
