 G'day all, thanks for the invite. Pleased to be here down in sunny Adelaide. So I'm here to talk about imagery and crop analytics in broadacre farming. So who are hummingbird? We're an advanced crop analytics business. So born out of London, we've been in Oz for about 15 months now. Most of our work's being done over in Western Australia on the cereals crop plus canola, but focusing a lot of work on the wheat and barley space as well. In terms of our team globally, we've got 64 in there, 40 of which are data scientists and engineers. So similar to Tim's business, we love data going through our business in terms of whether it's satellite or drone imagery. The more imagery that goes through, the better the algorithms get and the better the output to the farmer. In terms of, yeah, the countries we operate in. So we've got 250 customers spread across the UK, Russia, Ukraine, Australia and now Canada as well. So here's a snapshot of our process. So as I said, we're data collection agnostic, meaning that any imagery that comes in from satellite, drone or plane can be analysed through our system, which is where the IP is, and then provide an output depending on what the farmer needs. It can be as simple as an NDVI map or as complicated as a high res kind of weed detection map, which I'll go through later. And those shapefiles, when you obtain a prescription map, integrate directly into your farm equipment, whether it be your spreader or your sprayer. We did a funding round last year and a lot of that has been spent integrating with the equipment to make it easier for the farmer. So going through satellite, so NDVI, you've all heard it before. It's a common metric, but I think it's a real important one for broad acre crops. You've got these massive farms that I think the scale of them, it's not quite possible to scout them on a regular basis. So having the ability to be directed to areas and not take the agronomist role, but merely tell the agronomists and the farmer where might be an issue and have a better utilisation of time is really important. So I'll just take you through this. So in terms of the map itself, and I've got a couple of examples, just quantifying vegetation based on the infrared bands. So determining whether the crop's healthy, stressed or dead. The benefit is it gives you an overview of your whole crop and even your whole farm in an instant manner. So subject to cloud cover, which is going to change because there's technology that sees through the clouds. You receive an image every week. You'll then have a time series of images and be able to compare different images and see what's changed between certain dates, which is important. And as Tim alluded to with the variability in fields, it's so great across Australia and I don't think it's being tended to in the best possible manner and using the best means, which is these imagery analytics. So in terms of our NDVI, we use our own transformation and calibration rather than off-the-shelf software to ensure that the Sentinel-2 images reflect what's actually happening on the ground. And in terms of the farmers themselves, it highlights areas where they need to improve on, so whether that be soil structure, nutrition or even drainage. So again, alluding to the time management piece, so just having that snapshot aerial overview of the crop, if you compare it to viewing a 300 hectare field from the side, you've got 2Ks that way and a K.5 that way, it's pretty hard to see what goes on in the middle. So determining where the issues are quickly and then having the ability to go and fix them. I think it's also important with NDVI, it's a metric, so it's an index, so it's giving you that start point. So you can go back in time three years and overlay what's happened over those fields during those years, and then you can overlay other things like your yield maps, rainfall, temperature. But having that NDVI score as a start point, you're comparing x against x, which I think is important and takes out the human element of risk. So here's an example here of a farm over in WA. As you can see, the crops, it's not very green yet, meaning the crop hasn't emerged or it's not so healthy. It's been pretty dry over there as well, so this is from last season. Now this is an image seven weeks later, so what's good about the hummingbird platform or one of the advantages is you can, once the fields are uploaded, you can click a compare button. So then you're looking at the image up the top, which was taken on the 17th of August, and you can see down in that bottom left of the field there, there's something happening, so you can tend to it straight away, and then 10 days later you can see that image is, well that issue is no longer there. It could have been a weed infestation or anything really, and it's been tended to. And in terms of the adjustment bar, so you've got the NDVI scores, you can manually adjust that metric to focus on areas of concern. So you kind of, where the crops got a nice even coverage, you can then swing that bar to focus on the areas which need attention, and then get cross chest against the RGB, which is just your photographic image. So this is a variability of crop growth rate map, so this is looking at a time series of two different images, and just showing where it's growing fast and where it's growing slow in the fields, which I think is a really useful tool, particularly if you've got farms that are above 5,000 hectares or even 3,000 hectares, if you want to look at your whole field and determine which fields, or you're talking about prioritising harvest, you can look at that aerial snapshot of your whole farm and determine where it's growing faster and slower and just prioritise your harvest. So cloud cover is an issue, so at present it is going to be eliminated, I think there's technologies coming out that's going to pierce through the clouds, but we've actually developed a cloud masking algorithm. So when you're downloading your VR map or prescription map, it asks you whether you want to, how you want to spray those areas under the clouds, because the cloud does skew the NDVI scores, so I think that's important. Again, this is just a dataset where you have to compare the photographic imagery to what's coming up on the NDVI just to ensure that your scores aren't skewed. So in terms of our advantages, so we've got the photographic or RGB imagery, which is available to cross-check against your NDVI, which allows you to further investigate what the differences relate to. The adjustable slider, so you can put your own bespoke scoring to further investigate areas of interest, and then the cloud masking algorithm, which is pretty critical in terms of not skewing the NDVI scores on those fields. Now looking at our UAV products, which we've done a lot of work on over in WA and hopefully do a lot of work here in South Australia. So what this product does, it's a knockdown product. So we brought a budget in your summer sprays. It's aimed at reducing your summer spray costs. So at the moment, normal practice is to blanket apply your knockdown mix across 100% of the fields. What this does is you're flying a field with a UAV or a drone and a high resolution. So you're mapping the weeds that are kind of that cocaine size and above into a prescription map, and then it goes through our system and then it uploads to your boom spray equipment. So you're competing with your weed-seekers, but what we're saying is you've mapped the weeds before your spray goes out there so you can better plan in terms of what herbicide mix you're going to use. It also integrates with your current equipment, so you're not going to lose any efficiencies there in terms of speed and the like. And you've also got the prescription map as a data point, which allows you to ultimately overlay with your map to ensure you've got all the weeds. If you've got more than 70% coverage of weeds, you may as well blanket apply. So it's obviously not applicable, but it's when it's that 20%, 30%, 40% range where you're thinking, oh, do I bring in the contract sprayer? Or you've been thinking about buying a weeded. And we've assumed a lot of our projections based on a $13 hectare knockdown mix. I don't know what people use here, but I've heard it's kind of in the realms of $17 to $19 is what people are using, so it just creates a better ROI. So the fields are flowing using, so it's detecting green on brown, so it's picking up the green against the dirt. You can capture up to 1,000 hectares in a day. Once that data is captured, it's uploaded to our platform and processed, then gives you a prescription map, which integrates with your spray equipment. So at the moment, you've got to download it onto USB and chuck it in. But in due course, it'll be kind of drive past the bore, and hopefully you've got a remote sensing device, and it just integrates automatically. So here's some case studies. So this is a field in WA where we've, it was one of the first ones we flew. So we've, you can see where we've mapped the weeds there, the green areas where we need to spray. So using that solution, we only sprayed 50% of the fields as opposed to 100%, and there's the RGB or photographic compare. In terms of the ground truthing, we always go around the fields and map the different weeds as obviously different weeds. So they've got different signatures and the like. RGB doesn't, it doesn't matter with the signatures, but with the multi-spec, you know, if you're doing any weed mapping, you need to know what the signatures are. And in terms of the economics, so in this instance, it was a $13 hectare herbicide mix. Our cost is four bucks a hectare. So we're spraying 50% of the fields. The total cost is your $10.50 a hectare. So your return's $19%, or $2.50 a hectare on that particular field. In terms of integration, it really is quite simple. As I alluded to, a lot of our last bit of funding was allocated to the integration piece, which I think was a big blocker in this space previously. And so that's really assisted in making it as easy as possible for the farmer. So here's, I guess, a more commercial case study whereby we've flown a field that's 300 hectares, and only 44% of the field got sprayed. And what's even more beneficial is their herbicide mix was nearly $15 a hectare. So it resulted in a saving of nearly $4.30 a hectare, or 29%. So if you kind of extrapolate those figures across a full farm, it's a real material saving to your costs and hopefully yields to. That's it.