AI: Fly Strike Monitoring in Sheep

One of the biggest issues farmers face when it comes to sheep farming is Myiasis or more commonly known as fly strike. This occurs when wet, warm conditions attract flies to dirty, wet wool. They lay their eggs and when they hatch, they eat away at the host’s flesh (Fly Strike in Pets, 2025). This occurs quickly, with the fly’s life cycle only taking 10 days. While it is treatable, it is tedious and expensive and time consuming. Wool must be sheared and the affected area treated with insecticide or larvacide, and some sheep with antibiotics. Frequently, animals need to be humanely euthanized (Flystrike, 2025). This poses a huge economic threat to farmers, especially with large mobs of sheep which are difficult to spot the first signs of fly strike. These are often loose, hanging wool, wet wool, sores, and skin inflammation (often looking bare, dry and pink) (Flystrike, 2025). A possible solution could be to use Ai. By programming an app to recognize when a sheep shows signs of fly strike, farmers will be able to monitor their sheep for signs of this deadly plight more readily.

Ai already knows how to detect specific species and traits. It’s been used to identify specific fungal species:

“machine-learning species identifier has been developed that takes as input locality data and a small number of the morphological parameter” (Bartlett et. Al, 2022)

It works off of a data set. Therefor, a large data set of images of sheep in different positions, with different symptoms of the fly strike would have to be uploaded. Covering different breeds. Ages, sizes and colours as well. After that, an Ai model will have to be trained on the images and learn an algorithm to identify signs of illness. There’s two possible Ai training models that could be considered:

Convolutional Neural Networks. These are frequently used by security cameras, autonomous vehicles and other uses to identify objects. They have multiple layers, and “If the output of any individual node is above the specified threshold value, that node is activated, sending data to the next layer of the network”(Ibm, 2024). This approach provides a reasonable approach to classifying and identifying images.

Computer vision is a newer form of Ai. Not only can it identify objects, but make recommendations based on their visual input. (Shuresh, 2023). This means that not only could they identify the potentially at risk sheep, but suggest how imminent treatment is and what treatment is appropriate.

Whichever model is used, The model will need to be tested and refined based on accuracy in the real world. False positives and negatives would have to be used.

The next consideration would be a device or app to run the program. Ideally drones or cameras would be most successful. Drones could monitor sheep in the paddock, and cameras could be placed in high trafficked areas. the ability for farmers to upload images to receive AI-powered health assessments on individual sheep would also be beneficial.

In order to keep the app profitable, a subscription model will be offered for continuous access to the tool. Tiers can be added for anyone from individual farmers to larger operations with multiple paddocks. There also could be data insights which could be bought for an added fee. History reports, trends, and veterinary reccomended treatments. Perhaps even an option to download data to partnering vet clinics.

Analytics like predictive insights (like pairing with weather forecasting to predict when fly strike is most likely to occur) can be an added offer.

Possible expansion:

• Partner with veterinary services to offer a diagnostic tool. Vets can use your AI solution to monitor larger flocks remotely and provide more accurate diagnoses.
• Farm Equipment Manufacturers: Partner with companies that sell farming equipment (like drones, automated cameras, or sensors) to integrate your AI solution into their devices, providing a complete package. Larger operations may want more specific devices rather than an app.

Leverage AI-Driven Automation

• Automated Detection System: Develop automated systems that alert farmers when a sick sheep is detected. This could involve smart cameras or drones patrolling the paddock, sending alerts to a farmer’s phone or computer when sickness is detected, as well as for projections of timely care.
• AI-Enhanced Decision Making: Use AI to suggest treatment options, preventive measures, or even advise on optimal flock management based on the ongoing health status.

6. Offer Training & Consultation

• Provide training and consultancy services to farms on how to use the AI technology effectively. This could include setting up the system, interpreting results, and integrating it with their daily practices. Government grants may be a possible solution to integrating this technology.

7. Expand to Other Livestock

• Diversify: Once the system is optimized for sheep, you could expand it to other livestock, or even everyday pets like rabbits.

• Given the agricultural focus and tech innovation, you may qualify for government grants, subsidies, or investment from AgTech venture capitalists looking for impactful tech solutions in farming. Australia already has multiple grant offers for farmers wanting to upgrade their operation, such as solar power.

By developing and marketing an AI solution that addresses livestock health monitoring, you can tap into a growing market within the agriculture industry. It could significantly reduce costs for farmers while improving animal welfare and productivity.

Flystrike in pets – risk factors & prevention: RSPCA – RSPCA. RSPCA. (n.d.). https://www.rspca.org.uk/adviceandwelfare/pets/general/flystrike

Flystrike. Veterinary handbook for cattle, sheep and goats > Diseases. (n.d.). 2025 https://www.veterinaryhandbook.com.au/Diseases.aspx?diseasenameid=97

Bartlett, P., Eberhardt, U., Schütz, N. et al.Species determination using AI machine-learning algorithms: Hebeloma as a case study.IMA Fungus 13, 13 (2022). https://doi.org/10.1186/s43008-022-00099-x

Ibm. (2024, December 19). What are convolutional neural networks?. IBM. https://www.ibm.com/think/topics/convolutional-neural-networks

Suresh, S. (2023, July 5). Computer vision: A comprehensive guide to techniques and applications. OPIT. https://www.opit.com/magazine/computer-vision-2/