UAVs and Precision Viticulture

Dear UAV fellows,

now that I have a drone with RGB and NGB cameras onboard, Pixhawk working well (except airspeed sensor), automatic triggering of both cameras at the same time (sync), Photoscan Pro to deliver orthomosaic/DEM pretty easy and fast, I can focus on the applications. I've selected vineyards and wine makers and surprisingly I could not find any straightforward way how I can sell my product. So let's find out if something interesting comes out of the discussion here. If you have your own experience in this area, please share it here.

I have access to a local wine grower here in Slovakia. I want to use this opportunity to do my own research and share the results and observations here. I intend to go out and map the vineyard (30ha) frequently according to the weather. The grower has several types of grapes and the main purpose would be to find out if NDVI or other VI can be used to monitor the plant stress and the reason why it occurs. I've included some figures from the book Remote Sensing of Vegetation 2010 that could be a good starting point. I realize this could be a long way to find out the simple facts that could help the growers and that the research is ongoing already elsewhere. First I have to find some source of funding to do this long-term research..

20140701_134844.jpg

20140701_135305.jpg

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –

Replies

  • 3D Robotics

    Hi Michal,

    While we don't have any specific experience in viticulture right now, although as soon as we get a chance we are headed to the Western Slope of Colorado to work with a few vineyards, I think the lessons we learned with row crops as far as turning UAV images into a product will apply to you as well. First, I would fly as many vineyards and sit down with as many viticulturists as possible and learn everything you can about growing grapes. Second, I would identify the costliest inputs in grape production and see if your imagery can address any of these. Finally, I would select a number of farms to serve as pilot customers and, with their guidance, develop a value proposition that will help you price your product. If you determine that your imagery can identify a pest that can be addressed through spot application of herbicide that is found only on a certain chardonnay variety in a certain climate zone, then that becomes the seed of your business and you can market yourself specifically to those growers.

    We followed an identical process for row crops and determined specific value propositions in weed identification and, more recently, fertilizer prescription and early yield estimates. I'm certain that many other applications exist for our imagery, but quantifying input savings and increased yields for specific use cases will make the sales process much easier. The idea is that we turn a our toy into a farmer's tool.

    With regards to the differences between NDVI and DVI imagery I previously wrote about, it has become abundantly clear that for our collection techniques and application DVI images are far superior. I will try to post an update today showing some more ground truthing, but we have become very confident at this point that the DVI images we collect correlate well with the conditions on the ground.

    Best,

    Daniel

    • Daniel, if you can make it out to California, I can hook you up with a friend at the UC Davis experimental research vineyard in Napa Valley who would be interested - he researches the red blotch disease I mentioned earlier.  Just let me know.

    • 3D Robotics

      Hi Tim,

      That would be great. We've focused our efforts on developing a product honed for quarter section row crops, but see a lot of potential in viticulture and would love to work with an experimental farm. I'm not totally sure a fixed wing is the best platform for vineyards (I don't know much about grapes, but the ones I've visited as a tourist were generally less than 10 acres), but we definitely could apply everything we've learned about image collection and processing to a new problem. We'll also be out in the San Francisco area in early August so the timing could work out really well. If it's easier, my email is daniel at agribotix dot com. I look forward to connecting.

      Thanks,

      Daniel

    • Thanks Daniel !

      Please read my post above, I have some bad experience with the first agronomist who works for the local wine grower. They were not interested in identification/localization of diseases/pests because they say it's too late.. If I could detect the location before the disease appears, that could be of value for them. But I fear this is too much to ask of any spectral imagery.. It would also require daily flights with daily DVI maps and I don't have any source of funding for this research that has to take whole year probably. I thought that maybe someone could be interested to fund this but it's hard to find such source. Maybe different wine grower.. Maybe a grant application from some University.. It will take time until I find the right contacts to all possible targets within 100km range of my home.. I realize only a few might be interested in this technology anyway..

      Regarding the weed identification - please what type of main crop was in focus ? I suppose a grain, corn or wheat is not of interest here.. maybe tomato or potato ?

    • Dear Michal,

      I am very interested in your project.

      You are welcome to came to us to stay and to make all the tests that you need and to find the right

      utilization for the UAV in the vineyards.

      We are the biggest Romanian Premium wine producer.(www.jidvei.ro).

    • Wow, great ! We can discuss it in private now.

    • 3D Robotics

      Hi Michal,


      We use DVI images to quantify weed coverage on fallow fields, so the type of crop isn't terribly important. Organic farmers need this information to determine when to take a field out of organic rotation and conventional farmers need this to spot spray round-up resistant weeds. There is some potential to identify weeds in a growing field using different spectral properties, but we only have done this successfully with one weed type.

      Best,

      Daniel

  • Hi Michal,

    From a single NIR image I can already see the same differences that an NDVI is going to point out to be honest. The sensors in these consumer cameras capture a very wide band of the spectrum, which makes it harder to extract very specific information, you'd need multispectral sensors to be able to extract more information. NDVI is a bit like only having a hammer when you want to construct a house, quite a blunt tool.

    Selling to individuals costs a lot of effort in convincing and test flying while not making money. Focus on a decent sized organization first and run experiments with them. There was a story about McCain looking into drone use for potato crop monitoring. That is an excellent way to immediately push ahead for other customers and get the exposure you need.

    Not all viticulturists are concerned about their crop. Here 1/5 grows wine with noble grapes, the others are very regular wines really. The immediate concerns they have are "is there going to be rainfall this year?" and the other is "is anyone invading my vineyard and constructing houses?". Precision Ag only comes somewhere after that, so you shouldn't expect jaw-drops when you explain the technology to them, if you are in situations like that. Best is to take an open-minded approach to things, you may be able to sell services against invasion first and then supplement with other services. Always ask them "what's your biggest concern right now?" and that may give you pointers into new business.

    • Well I don't expect that consumer grade NIR sensitive camera will be the ultimate solution here. It is just the starting point that is cheap and will point out something you can't see in visible light. I'm not a millionaire who can buy anything and this is probably the same situation with most of the users here. As I wrote before in this blog, hyperspectral camera is too expensive just for testing. I know it starts on a price tag about 60-75000 EUR and that is the cheapest sensor with who knows what quality if you cannot test it.

      I don't know anything about thermal cameras and their price range, but certainly it's not something cheap and the resolution for the vine might be too low too. But I will investigate in this matter..

      Serious multispectral cameras like Tetracam that are light and cheap enough might be an option, but again the bands are broad and the result might be similar to consumer grade camera that is even cheaper. So why to invest ?

      I'm not selling to individuals with small vineyard. My first potential client has a 30 hectare field with several types of vine and he is profit oriented with lot's of technology already invested. They have own agronomist and I already had a meeting with them. It is the nearest vine grower who could be interested if I have something to show. So it's now a question if I want to invest a time to frequently fly over his vineyard and observe some pattern in NDVI / VI data with consumer grade camera or rather invest into different sensor/camera that would help me get out the information I can sell better. But my financies are very limited and I don't want to go to bank and borrow money just to find out that at he end I have no product to sell. It's a huge risk for me but that is the essence of business right :) First I want to study all literature about the topic, then see what could be the best option for this market.

      I don't have big eyes. Maybe vine growers are not the market for me after all. But farmers with other crops might be in the same situation. It's the technology that should make the difference, after all why are we doing all this ?

      I've as many others spent too much time and money to build a flyable drone with some cameras onboard. Now is the time to find out if there is a market to sell it somehow. I did not spent time on this just for fun.. If this industry is meant to be used in precision agriculture as the main source of income (according to many AUVSI talks or any market predictions I've seen in the last 5 years) then let's find out if that is true. Do you have some experience with this ?

    • Hi there.

      I came across this post randomly.  I am a long-time contributor to the AeroQuad project, and have only occasionally snooped around on DIYD, but this post caused me to register and comment.

      Im a US winemaker with a Viticulture & Enology Degree, as well as a Quad-Nerd.   I've thought a lot about viticultural applications for quads, and I have two suggestions for you, only one of which might play to the strengths of what you have developed.

      Vineyards in the US are fighting a vine virus called "red blotch"  - which surprisingly enough, causes red blotches on the leaves.  We believe the virus is being vectored by some kind of insect, but are not sure.   Either way, there will be interest in ways of quantifying the degree and the spread of the incidence of red blotch in vineyards both throughout the season, as well as from season to season.

      The other application that I see for autonomous craft in this industry is a lot more basic:  scaring away birds...

      If you can have a quad that flies over vineyards several times a day making sounds like a hawk... then you will have a lot of takers...    If you can make it dock like a roomba, charge, and go out on it's own, then you will have even more...  Not that such an idea plays to the strengths of your craft.

      The other applications for imagery such as mapping vine vigor, water stress, etc...  All of those things are something that you do a study on every so often so that you can change your watering zones and times... etc.   Something that someone would buy one of your vehicles and then provide a service to vineyards.  I dont see you selling a lot of vehicles for that purpose.

      That is my two cents.  Feel free to contact me directly if you want to chat on some of these topics.

      And remember: Corks suck.  Winemakers who care about quality use screwcaps!

This reply was deleted.