NZ Uses AI to Win Crop Wars
- By Lachlan Colquhoun
- June 18, 2018
New Zealand’s agriculture sector has ranked biosecurity as its number one issue and is now developing Artificial Intelligence (AI) solutions to help.
University of Canterbury academic Dr. Varvara Vetrova is working on AI and Big Data solutions that can identify suspect plants, fungi and insects from photographs.
Dr. Vetrova is a lecturer in the University's School of Mathematics and Statistics and is leading a three year, NZD 1 million research project.
With her colleagues, she is researching how AI and machine learning technologies can be used in protecting NZ agricultural businesses.
Called Biosecure ID, the project harnesses a machine which then automates the image-based identification of species based on “deep convolutional neural networks.”
“We are developing a prototype image-based system for rapid, automatic identification of potential pest species in New Zealand," Dr. Vetrova said.
The NZ Government is supporting the research through its Endeavour grant series, and the project involves researchers from the fields of biology, computer science, computer vision and applied machine learning.
Biosecurity has remained the highest-ranked priority for the New Zealand primary sector for the eighth year in a row, according to KPMG in its latest issue of AgriBusiness Agenda.
The trend is up for New Zealand's primary industry exports, but the risks are growing, rendering biosecurity a leading priority.
Primary industry exports are forecast to have grown 11.8% in the year ending June 2018 to NZD 42.6 billion, NZD 1 billion more than forecast, according to MPI’s latest Situation and Outlook.
"Unsurprisingly, against a background of Mycoplasma bovis and myrtle rust, world-class biosecurity remains a number one priority ranking for the eighth consecutive year, hitting a new record priority score of 9.62 in the process," it said.
Another digital issue was ranked second, with high-speed rural broadband hitting a priority score of 8.73.
On the biosecurity research, Dr. Vetrova said the aim was to build an app which could immediately identify specimens from photos.
Plants, fungi, and insects would be automatically identified in detail. Similar looking specimens would be easily distinguished.
“Provided that we have a snapshot of an organism, we could tell which species it is,” Dr. Vetrova said.
“So, a biosecurity officer or a farmer could find a bug, fungus or plant, take a picture and that the program can quickly tell them this is a worry or this is a New Zealand species.
"We are trying to build a targeted model; the main idea is that an interested group, for example, farmers, will ask to help identify particular weeds. Then they will supply us with a few images, and we will train and provide a model for them."