Modern Day Pasture Measurement with AIMER Vision

In our last article, we explored the untapped potential of better pasture measurement and management, as highlighted by Beukes et al. (2018). Their research showed that perfect knowledge of pasture mass and management could boost farm operating profit by $640 per hectare at $6.33/kg milksolids. That’s a massive number! But it got me thinking—how do we get 'perfect knowledge' of pasture mass and management?

Perfect vs. Low Knowledge: What’s the Difference?

When I think about ‘perfect knowledge’ and this is aligned with the simulation approach of Beukes, it comes down to a few key things:

• Accurate assessment of pasture mass and correct order of the longest to shortest paddocks to graze.
• Identifying the very day cattle are going into paddocks with too high pre-graze mass while maintaining the perfect rotation length i.e. for perennial ryegrass dominant pastures, the time it takes for ryegrass to hit its three-leaf stage.
• Realising when cattle will leave behind more than 1700 kg DM/ha after grazing and then doing something about it. That might mean cutting silage, reducing supplements, or adjusting daily plans to make sure residuals stay under 1700 kg DM/ha.

By contrast, ‘low knowledge’ means you’re:

• Guessing paddock mass.
• Only conserving pasture when post-graze mass is over 2000 kg DM/ha – you’ve lost control
• Feeding supplements on a flat daily schedule, no matter what’s actually happening in the paddocks.

Perfect knowledge isn’t easy to achieve, but as a pasture farmer, it should be the goal to maximize profits. That means any digital tool should make that goal its north star.

What Should a Digital Tool Do?

If I think about the goals of a pasture management tool, especially in a surplus situation, here’s two on my list:

1. Estimate paddock mass accurately when and where I need it (“estimation problem”).
2. Spot when pre-graze mass is too high—now or in the future—based on perfect rotation length, animal demand, pasture growth rate (“biophysical problem”).

But before tackling the biophysical problem, we need to get the basics right. So, let’s talk about measuring pasture mass – the estimation problem.

Measuring Pasture Mass: How Accurate Is Accurate Enough?

Here’s the reality check: We’ll never hit perfect accuracy unless we want to throw probably hundreds of millions of dollars at it—and even then, the economics just wouldn’t make sense. So, what’s “accurate enough”?

For me, it comes down to precisely allocating pasture to cattle each day, aligning with their nutritional needs to achieve the optimal residual. Research from 2017 by Callum Eastwood and Brian Dela Rue of?DairyNZ shows that farmers’ opinions vary:

• 40% say being within 250 kg DM/ha is good enough.
• Another 40% want to be within 125 kg DM/ha.

I’d argue that for pre-graze mass, we can accept less accuracy—say, within 200 kg DM/ha. But for post-graze mass, we need to be within 100 kg DM/ha to truly understand if we’re undergrazing or overgrazing to maximise pasture quality and regrowth.

What Makes a Great Estimation Tool?

After talking to farmers, looking at research, and thinking about my own experience, here’s what I think an estimation tool needs:

1. Ease of use: 76% of farmers say this is critical. Honestly, what could be easier than not measuring at all? That’s why satellites initially caught so much attention.
2. Weekly updates during high-growth periods: 75% of farmers want this. Make sense you want to get ahead of surpluses.
3. Rapid results: 71% of farmers say they need data within 24 hours. But let’s be real—if you’re deciding how many breaks to put up or whether to keep cattle in a break longer, you need data in seconds, not hours.
4. Low maintenance: 68% of farmers value this. Fair enough—no one wants more hassle.
5. Speed: 60% of farmers want data collection done in 2-3 hours for 100 ha. In the middle of calving, though, I’d say an hour is all I’ve got to spare.

Why a Smartphone?

This got me thinking: What device could meet all these needs? The answer seemed obvious—a smartphone. I already have one, it’s always with me, and software updates handle maintenance.

• Ease of use: Tick
• Availability: ?Tick
• Low maintenance: Tick
• Cost: Tick

Then I wondered… with machine vision advancing so quickly, could a smartphone camera estimate pasture mass? Could it do it fast enough? Accurately enough? That’s what we set out to figure out.

Bringing AIMER Vision to life!

Fast forward through countless data collections, model iterations, and tests, and we’ve now built AIMER Vision, AI enabled pasture measurement. Here’s what it can do:

• Estimate ~90% of pasture covers within 200 kg DM/ha of measured values for unseen data.
• Deliver results within 10 seconds after hitting record—even offline

It’s easy to use, always available, requires no maintenance, and works rapidly. For 2025, this is a game-changer in pasture management—and we’re not stopping here! We’ve got more extensions lined up for dairy and beef farmers worldwide.

What’s Next?

In the next article, I’ll share how we tackled the "biophysical problem" I mentioned earlier. Stay tuned!