nextGen Agri logo
nextGen Agri logo
Join / login

we are working on.

Innovation is at our core. We love learning new things about agriculture and applying new thinking to old problems. We regularly conduct research and development projects on behalf of industry organisations. We are very proud of our close collaborations with a range of leading industry organisations and universities across these projects.

Sheep farming at the edge

neXtgen Agri is focused on applying machine learning and machine vision to the Australasian sheep industry.

We have collaborated with several organisations and universities to build and analyse some of the largest agricultural data sets created for the sheep industry.

One example is a 1.4 million image data set used to train a sheep recognition classifier. Our current research involves:

  • Training classifiers to recognise individual sheep
  • Training multiple object trackers to track sheep in video
  • Developing systems that enable both training and prediction on edge devices like Google TPUs
  • Training classifiers to predict sheep behaviour from tri-axial accelerometer data
  • Training classifiers to predict the date and time of parturition
  • Training classifiers to recognise sheep behaviour from video
  • Integrating camera and machine learning hardware into remote farm-based monitoring systems
Job opportunity
Using smart tags to monitor sheep behaviour
Flock of sheep grazing in a green field

Footrot free fine wool.

On behalf of The New Zealand Merino Company (NZM)we are working toward a genetic solution to footrot.  

We are developing a footrot breeding value that can be used to reduce the susceptibility of sheep to this debilitating disease. This project has been ongoing since 2012 and involves the Animal Genetics and Breeding Unit (AGBU) and Sheep Genetics.  

The project has required significant collaboration with fine wool breeders across New Zealand.  We run the NZM Central Progeny Test as part of this work program. We are making some serious inroads into finding footrot resistant fine-wool sheep.  For more information contact us, or visit NZM's Perfect Sheep website.

ZQ Report

Grazing Bytes.

Within the Grazing Bytes project we are investigating the ability to assess the amount of pasture that sheep are consuming out in the paddock, in real time. 

The project is using Smart Tags that have been developed by Australian Wool Innovation (AWI) to undertake this work. If successful, the results from this project will enable producers to understand what is happening out in their paddocks with a lot more accuracy and allow for more targeted nutrition and more efficient pasture utilisation. 

The project is using a combination of old science and very new science in the form of machine learning.  We are hoping to achieve what has been thought of by many researchers in the past as the holy grail of sheep research –accurately assessing pasture intake of grazing sheep.

This is a collaborative project between Australian Wool Innovation, neXtgen Agri, Murdoch University, Muresk Institute and Agriculture Victoria.

Australian Wool Innovation LogoAgriculture Victoria LogoMurdoch University LogoMuresk Institute Logo
Flock of footrot free fine-wool sheep in a green field
Close-up view of a sheep fleece

A new strategy on flystrike.

We are currently undertaking a review of flystrike and proposing a research strategy for future research on this important topic.

We are conducting this work on behalf of Meat and Livestock Australia. 

To carry out this work we have brought together an impressive line-up of flystrike gurus – Professor Herman Raadsma, Professor Bill Pomroy, Dr David Scobie alongside our own Dr Mark Ferguson. 

This has given us the opportunity to not just review what has been done to date but to do some blue sky thinking on what the focus of future research efforts could be.

Meat & Livestock Australia

Behaviour Monitoring.

The ability to detect animal behaviour using sensors has started to become a reality over the last few years. 

In order to be able to use this new capability, the industry needs large banks of data that can be utilised to train algorithms. Within this project we are building the world’s largest sheep behaviour data bank. 

This involves over 500 hours of video of sheep grazing, walking, sitting, standing and ruminating across a range of grazing scenarios. This footage is then meticulously examined to determine the activity of each of 10 sheep in 10 second intervals. 

The sheep in the study have been fitted with a range of sensors. The sensor data is combined with the sheep behaviour data and this information is used to train a machine learning model. 

This project is a collaboration between Meat and Livestock Australia, Murdoch University, neXtgen Agri, Muresk Institute and the Western Australian Department of Primary Industries and Regional Development.

Using smart tags to monitor sheep behaviour

Like our content? Want to know when we post something new?

Join our newsletter and we will send you a email from time to time to let you know when we have something new for you.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.