Over the last few years, farming has come under fire for doing more to harm the planet than to heal it.
The dairy sector, in particular, is experiencing problems, such as unsustainable dairy farming and inadequate fertilizer handling, that are escalating the climate change problem and which need resolving ASAP.
Technology comes to the rescue in the form of CattleEye, an AI-first cattle monitoring company that partnered up with London-based AI data engine V7 to develop intelligent cattle monitoring solutions that can enhance cattle’s welfare, reduce carbon footprint and help combat farm labor shortages with automation.
The Tech Behind CattleEye
CattleEye works with trained vets to build accurate computer vision models that can detect lameness in cows and other health issues as soon as possible.
To date, there are about 20,000 cows monitored by CattleEye’s systems on farms throughout the UK and US, and the farms are benefiting from extended cow production cycles.
It works like this: CattleEye collects video data from numerous farms. Then, the team collaborates with experienced vets who annotate data on V7, labeling cows as they walk past the cameras, get milked or eat, and highlighting anomalies (e.g., lameness) or pregnancies.
The team also assigns mobility scores to understand cows’ health and performance better. This then allows the system to instantly flag any anomalies, such as poor mobility, whenever they occur once the technology is in use.
Annotated video data is then loaded onto V7, where vets would begin distinguishing different cows’ behaviors and use it to train an AI model that is installed in the cameras located on the farms.
The cameras monitor the cattle, capture their movements, and inform the farmers about any potential health issues or cattle behavior anomalies. The farmer can then either take immediate action or keep a close eye on the cow.
After a few hundred examples are provided, the teaching is sped-up as early AI models begin pre-completing training data to cover cases where their confidence is high.
Going from on-farm visits with vets manually recording data to fully remote scoring of videos in V7, we were able to greatly enrich the variety of training data from farms across the world. This also saves the busy vets travel time in addition to reducing the potential impact on the herd as cows do not need to be disturbed with on-farm visits.
Ryan McMillan, Lead Data Engineer at CattleEye
The more data collected and annotated, the more accurate the model becomes. CattleEye’s AI can spot even the slightest movements of every cow, proving that AI can replace manual animal matching and on-site vet visits, thus saving an average farmer £350 per cow per year.
Using AI for combating carbon footprint with smart farming
Farms can save thousands of dollars thanks to the technology developed by V7 and CattleEye, which proves that AI can do the same job as farmers when it comes to detecting cow lameness – but better, faster, and cheaper.
Not just that, but Terry Canning, CEO of CattleEye, points out that
In 50 years’ time, 90% of human labor on farms will be replaced by machines.
This is ideal for farmers who currently work a backbreaking 15-17 hours a day, as well as farms that are struggling with the labor shortage, and farmers who simply see detecting cow lameness as a time-consuming chore.
However, as awesome as all this is, the company is driven by a bigger vision: Reducing the carbon footprint.
It’s an important vision, with total emissions from worldwide livestock currently amounting to 14.5% of all yearly anthropogenic GHG emissions.
And because cattle represents around 65% of said livestock emissions, it’s natural that CattleEye would set their sights on this particular sector of the dairy industry.
Monitoring cattle’s health and behavior, before taking immediate proactive action to prevent any health issues, is at the core of how CattleEye helps to reduce our global carbon footprint.
According to CattleEye’s CEO, Terry Canning:
We’ve actually calculated that if you can reduce lameness levels by 10% on a farm, there’s a saving of half a tonne of carbon per cow per year.
This is because reducing lameness levels boosts a cow’s efficiency.
Indeed, CattleEye is on a mission to reduce carbon footprint by 30% by 2050, with their machine learning system catching the eye of important investors who share their vision.
Training better AI with humans-in-the-loop
Artificial intelligence emulates human intelligence by learning from examples. For an AI model to learn to detect a cow, recognize its movements and assign it a mobility score, it must learn from a lot of human knowledge.
How does it do this?
This “knowledge” is called training data. The process of humans “teaching” AI by producing training data is known as annotation. This involves drawing boxes or shapes around objects on software such as V7 and then classifying the “item,” such as a cow, as a certain health condition, or giving it a mobility score. The more training data there is, the better the algorithm.
Once enough training data has been collected, AI models can be trained to ingest this data and test themselves against un-seen images. What has set V7’s technology apart is the ability for humans and AI models to work side-by-side on annotation challenges.
Reliability is also a challenge.
Whilst most engineering disciplines have predictable failures, AI models can’t be “looked into” to analyze what they don’t know. This is why understanding an AI’s training data is so crucial, because it acts as the repository of that AI’s total knowledge.
Once an AI is “trained”, this knowledge is compressed into a neural network’s millions of values, which aren’t readable by humans.
The future of AI in farming
With the United Nations suggesting that the world’s population will rocket to 9,700,000,000 of us by 2050, the stark truth is that the agriculture industry as a whole will have to double its current production levels.
It’s a huge task, but ultimately technology like CattleEye can help farms become more profitable and efficient by keeping a close eye on livestock’s health, monitoring them each day for their food intake and activity levels.
It’s not the only piece of technology-driven by machine learning and computer vision that’s available, too. Other systems have been created that can help farmers with the inspection of their plants, the fertilization of their crops – and much, much more.
In fact, it won’t be that long in the future when a farmer is able to carry out most of the work from the comfort of their kitchen using just a smartphone device.
V7 is a London-based AI startup founded in 2018. It develops an online training data platform to automate the process of AI data annotation, enabling AI-first companies to programmatically add labels to images or video. Its customers include GE, Boston Scientific, Fujifilm, and Miele. V7’s mission is to enable any business to create AI engines that automatically solve any task by emulating human actions performed on their platform.
Founded in 2019, CattleEye is on a mission to overhaul the dairy sector and reduce our global carbon footprint via technology that seeks to “unleash your cow’s potential,” thus boosting cow efficiencies. The technology is hands-free, with the company working closely with expert vets to train algorithms that quickly and accurately detect cow anomalies that can then be immediately rectified.