Tell me more: a deeper dive into the FARMER'S DASHBOARD

<p><strong>The following video describes the functions of the ROMI farmers dashboard and its potentials for farm management and analytics. Funded by EU Grant 773875.</strong></p> <p><em>Videos are available in:</em></p> <ul> <li>Hi-res (1080p Apple ProRes)</li> <li>Mid-res&nbsp;(1080p&nbsp;H265)</li> </ul> <p><strong>Video script:</strong></p> <p><strong>(HANAPPE) </strong>The farmer&#39;s dashboard is a tool, where it&#39;s still not in its final form as it is now. Right now it&#39;s tailored to the work we do in ROMI, where we collect the data from the cable bot and have like overview of what&#39;s growing in the field, extract you know, how individual plants are growing, how well they&#39;re growing, and in the longer term it becomes part of a bigger farm management system where we&#39;ll combine it with you know this the planning of what will be grown and when. And for these small farms they they do several harvest during the year they have weekly vegetable baskets, so they need some planning upon what they will seed when, when they will harvest some of the crops what will be the composition of the vegetable baskets that they&#39;ll be selling. And using these statistics that we get, we can try to help them adapt or predict a little bit more, the harvesting and then and what they can be, and what they&#39;ll be selling.&nbsp;</p> <p><strong>(BAORI) </strong>Now this is not something that you&#39;re going to have on a laptop that you&#39;ll be walking around the farm with. It&#39;s something that you&#39;re more going to use for analytics at the end of the day. So if you think about a situation perhaps where you will be doing daily scans say on your crop, or weekly depending on the size of the crop, and how many rovers you&#39;ve got and all that kind of stuff. You&#39;re going to be taking pictures of a whole bed and then at the end of the day that information will get uploaded to a computer that will then run some analytics and some metrics on what you&#39;re growing so you can have a little look and see where the areas for example of your crop are most productive, whether least productive, do you have an infestation, do you have some kind of disease in your crop, or is it just that it&#39;s very cold or something in that kind of, that area. And the crops are not responding as well as the rest. So once you start to put in place the scans and this information you type back to the rover&#39;s dashboard. It&#39;s all presented in a really easy kind of understandable presentation view, where you know anybody without even technical skill is going to be able to have a look and say oh there&#39;s a problem here.</p> <p><strong>(MINCHIN) </strong>So this parcel each parcel has a number and a geo reference which corresponds to the farmer&#39;s dashboard. And that means that we know both in space but also in a computer system what is planted where. This really helps when we&#39;re working in the field and we just want to know what&#39;s been done and what should be done, and then logging that back into the system. So just having a reference number allows us to be able to work in more detail</p> <p><strong>(HANAPPE) </strong>It becomes also I think an interesting data for research, so this data of growth of plants combined with weather data and so on. We can share between farms we can improve statistics so we can have like a better idea of what varieties grows well in what sort of conditions and so on. It is still open open question because there&#39;s of course a lot of parameters, but trying to start from something very concrete, potentially optimising the planning a bit over time we may be able to get some ideas about what varieties are well adapted in what regions and so on, and fine-tune how long it takes between seeding and harvesting and so on. So i see the farm bots evolve in that sense as a farm management system that combines data from the field like the data we get from the cable bot. That combines a planning that the farmer makes in the beginning of the year and potentially some modelling about how plants grow using these statistics that we&#39;ve collected over the years.&nbsp;</p> <p><strong>(COLLIAUX) </strong>For the farmers dashboard we have plenty of pair of images where the images correspond, too close by position of the camera of a cable bot. And we want to know how what is the displacement of the camera between the two images so that we can adjust the images to stitch them together. And we do this for many images along the path of the cable bot, and you see here we can generate a map of the of a culture bed. And then we can generate many maps every day of the culture bed. And we can see the radishes and actually the weeds here grow along the week and then we detect each of the plants and we mat match them along days, so that we can grow the growth curve of each plant and gather statistics about their growth. So you see here occurs with the statistics for the growth of plants and we hope, as I said before, that this kind of information is useful for farmers to predict their harvest in the coming weeks and also to design and manage their field.</p>