|
|
| Good afternoon!
I like to lurk around Agriculture forums to absorb as much as I can, and I was excited to see that you guys have a very active drone forum.
I'm not here to advertise or pick up business, so I'll omit the business name...but I own a drones-as-a-service company that's a little over a year old. The other two owners are an Agronomist and precision agriculture instructor, as well as a commercial pilot. My background is in networking and communications. Our main "workhorse" is the Micasense Rededge3, but we do have a bucket full of converted consumer cameras as well. We transitioned from custom built drones to commercial drones about 5 months ago, for the sake of easily training new operators.
Even though our Agronomist is definitely the POC for agriculture projects, I like to absorb as much information as I can manage. That's the main reason I'm here, but I thought that I would introduce myself on the off-chance that I'm able to answer some of your questions as well.
Thank you! |
|
| |
|
north east IN | Welcome to the group Derrick. Hopefully You will have a wealth of knowledge and experience you can share with us |
|
| |
|
| Thanks David. I don't know about the wealth of knowledge part, but hopefully I can offer up something! |
|
| |
|
southern MN | I suspect several of the fellas on here sell drone stuff.
This site is pretty strict about advertising or self promotion. Once they travel down that path the floodgates of spam open up.
You did well with your comments and control.
I suspect having a good email in your profile can get some good contacts, without breaking any advertising rules.
Welcome to the place, just as this section of the web site opened, my drone experience went off kilter, so I have been in a bit of a winter funk on it all. Bad batteries, bad updating support from the most popular drone co, too cold to fly, and now ever lasting 30mph winds.....
Paul |
|
| |
|
Des Moines, IA | I'm actually looking at getting into a UAS career myself. I've been building as a hobbyist for about four years now, but am looking to roll that into something more. I currently work for a large ag company, but have been looking for an opportunity to present itself for transitioning into something with drones.
I'm curious if you were part of the initial growing of your company or if you came into it later on. I'd also be interested in the services you guys provide to farmers. It is a daunting field to try to break into as a dedicated hobbyist. I've been trying to do testing and working on my processes on my parents' farm when I can get up there, but would love to get insight from someone actually working in that area.
If you'd rather not on here, I can shoot you my email and go from there. |
|
| |
|
| Thanks Paul. I sell a service, not a product, so it would make even less sense for me to try to sell it here (if I were so inclined). I'm definitely not a salesman and often get told that my conversations with potential customers about drones sound "counter productive", so I hope it'll be obvious after a time that I'm not trying to drum up business here.
A few things that I preach all the time are:
1. Precision agriculture doesn't NEED drones, and they do not currently fit into most workflows as a diagnostic tool. We'll provide the service if it can be used, but it often makes more sense for the grower to just get their own and take pictures from the air.
2. NDVI has become a buzzword in the drone world that means "magic wand", instead of an acronym that has a deeper meaning and some specific applications. It's not a new service or technology, and it's not applicable to all crop varieties and environments. It's also hard to generate accurate results, and you lose the ability to see obvious signs of poor data collection once you hand the pictures off to be processed. Ok anyways, I'll cut myself off before I go into my normal NDVI rant.
I hear you on the weather issue though. We have about a 10 hour drive planned for Friday and the weather looks iffy. Too marginal to cancel, though. |
|
| |
|
| My personal opinion is that the people that stand to make a lot of money in this are the software developers and hardware manufacturers. Service providers (us) are likely to be replaced fairly quickly, since the entire appeal of the industry is automation. They'll get much easier for your average Joe to operate and collect data, and the cloud services will get more reliable in their accuracy checking and analytics. I wouldn't be surprised if we end up with large endurance drones that cover entire regions weekly and offer the imagery as a subscription service as well, like satellites. Who knows.
I actually started the company by myself. I made a career change about 5 years ago and am a full time firefighter. That's why I became interested in drones. I got a lot of interest from farmers when they saw what cameras I was mounting on them for firefighting, which is where the whole agriculture angle came from for me (and how I met the Agronomist that is now part of the company).
Due to the experience needed to accurately interpret the data (and the potential liability), all we do for individual growers is provide the acquisition and processing, as well as the GIS work and index creation. I don't interpret the results. I actually talk a lot of people out of the multispectral outputs right now, unless they have a good understanding and are familiar with it. I think that the industry needs a couple more years of trending and analysis to be able to automate the interpretation of some of the data and establish standards of practice. Until then, RGB imagery provides a clear service for people that aren't familiar with the indices.
I prefer to do work for research driven projects, such as with universities, software developers and sensor manufacturers. Luckily, this is most of our work so far this year....it's also much steadier work, since most of them want flights on a bi-monthly basis over test plots, and occasionally more often for varying weather conditions. In these situations, we're really just the licensed and legal "field labor" for the people that are putting the real work into making the service useful. I'll support them as often as possible.
It's definitely daunting to break in to, but I'm probably the opposite of most of you. I had zero experience in agriculture 2 years ago(and still not much), but the rest of it was easier for me to dive into.
And I'm a big fan of sharing and learning publically, so I don't mind keeping it public (unless it needs to be private). |
|
| |
|
north east IN | Looks to me the best thing you can do with the drone is take high Quality pictures of your crops and tell a lot of what needs to be changed from that. I have seen a lot of problems that I have fixed from the pictures. The ndvi thing looks nice but what does it give you. I like the looks of the thermal and want to do some testing on that. Am I all wet on this. What is your thinking |
|
| |
|
| If multispectral imaging can't (reliably) deliver on the claims, then why over complicate the job by doing anything other than RGB? There is value in RGB, and you can avoid the uncertainties of the index maps by using something that our eyes and brains are used to making sense of. I'm kind of playing devil's advocate here and there IS value in the various indices, but they need to be carefully acquired and created and used in the correct manner.
I feel like we're in the stage of trying to finely tune the acquisition part so that the data interpretation software is able to offer assessments. I don't really think that automation is possible until you can have some sort of consistency with the data that's being plugged into it. I feel like there is a new camera announced every other week that is starting to address some of these issues, by having a built in IMU and using light sensors. Luckily, they're becoming much more affordable too.
The converted cameras are cool and definitely stuff have their use/place, but I feel like it takes so much work after the flight to try to prepare the images that it's just a hassle anymore. Some cameras now can spit out pictures that let you avoid a lot of that work, others are making progress in that direction.
NDVI is advertised as a "crop health map", but is primarily tied to LAI. It can be used in combination with other indices to try to identify specific traits or deficiencies in certain crops and certain soil types, but it's not quite to the point where it can be done automatically. I know there is a company in Europe (Germany?) that gives estimated biomass and nitrogen content results from drone flights, and they have restrictions on the conditions and cameras that they'll process. That's a good sign, because they're normalizing as many variables as possible before giving those rates. I have no idea whatsoever if they're accurate or not, but it's a good sign to me personally.
That's a long way of saying that yes, indices like NDVI can be useful in certain situations, but normal pictures have a clear value in most situations. Research projects aside, a good bit of our jobs are just maps made from these.
Our thermal camera is just set up for a live feed and firefighting right now, I've never mapped with it. There are some decent opportunities outside of Ag for them, but i'm pretty ignorant to the uses in Ag. If I remember right about what our Agronomist wants to test them for, it was something to do with irrigation, as well as soil temperature variance throughout the day. Not sure on that one. |
|
| |
|
Eastern Shore of MD | Welcome, Derrick! Reading through your posts so far, I think you'll be a valuable addition to this forum!
In regards to value of NDVI, I think its value is that it can quantify what we can (sometimes) see with RGB images and in addition things we can't see. Being able to make nutrient and other recommendations based off of them is something that I see in the near future, some are probably already doing that actually. As was mentioned by another poster, I have learned a lot by just stitching together some RGB images. I think that shows the potential of what NDVI can do. We just need to keep exploring that potential.
After some conversations with a couple people in the industry, they have raised concern to me about the quality of the images being taken by modified 'NIR' cameras. Their concern was that these images, in addition to questionable processing, can produce what LOOKS to be a good crop health map but the reality is that they are not entirely accurate. Every time I think I find a path I want to pursue with this, I learn more and, essentially, learn there is more I need to learn before jumping into it haha.
As far as the thermal imaging is concerned, my understanding was that its advantage was that it will catch stress in a plant before it actually begins to damage or reduce the health of a plant as would be seen in a NDVI image. Its an early warning to an early warning, if you will.
Edited by BacNBlak 4/27/2016 21:18
|
|
| |
|
| BacNBlak, your comment of - "Every time I think I find a path I want to pursue with this, I learn more and, essentially, learn there is more I need to learn before jumping into it haha." is the perfect attitude, I'm in the same boat.
Thanks to firefighting, I go through bouts of insomnia more often than not, and I might be a tad OCD. ;) After spending every waking moment reading about NDVI for months from every drone-related site I could find (and having actually done some testing), I learned more during a 20 minute phonecall with a remote sensing expert at the USDA than I had in the entire ~4 months previous. I was blown away at the amount of publically available white papers on the viability of multispectral sensors, both drone mounted and on manned aircraft. I couldn't believe that I hadn't stumbled on them before.
If I can find the bookmark, I'll edit and post them here. Some of the most applicable ones were testing a converted S100 and a big-name, calibrated rig over the same test plots at the same time. They also had RGB results from the same timeframe. The best place to get unbiased information is from the unbiased sources.
I'm definitely rooting for the drone industry, but I learned that it's not where I want to be getting my "education". Too easy to mix up sales pitches with facts. |
|
| |
|
southern MN | I agree the future is in subscription services from county wide drone services, all prepackaged data plan for you for $2, $5, $12 an acre depending on what level you want to buy into.
I'd wish we could do it ourselves with our own equipment and our own data, but the Information Age says it won't be so. That same data set can be packaged many different ways and sold to many other groups, marketers, other countries, buyers, interested parties I can't even think of... Its too valuable for us to keep to ourselves.*
Chewing through all the data generated, and figuring out what level of quality gives good info, is what we are going through now. Any reflected light depends on the sun strength, angle, cloud level, even altitude. So many variables to use that one pixel mean something uniformly across a day or season.
I hear thermal info can relate to plant stress, and gives us a 'view' of the plants we can't see with RGB? Figured you would be all over that.
Paul
|
|
| |
|
| Sorry, I forgot to touch on the thermal part.
With ALL of this, the "...catch x early..." comes down to how long of a window/early of an indicator do you have to catch it? Purely from an economical and operational viewpoint, I feel that it's not practical to fly all the property once a week in the hopes of maybe finding a problem early. This is based purely off of my experience as the person sending out the bills after a flight, not from any experience in farm operations whatsoever. I can see where it would make more sense for small area/high value crops, but that's assuming you're able to get good data out of that type of crop. I'd be curious to see a list of the most common signs and symptoms of these in different crops, and well proven early indicators with timetables attached.
Identifying invasive species is something that makes sense to me, because I can imagine how that process would take place on the software side and what indicators you're looking for in a big time window. Some of the other problems (I hear about rust alot), I've just never really heard a clear explanation of the progression speed and the associated signs, or how they could be identified from the air. I've been lucky enough to work on some projects with people much smarter than myself, so I'm relying on them to figure these things out, haha.
BacNBlak, do you happen to know exactly how the thermal/lwir is supposed to catch stress in a plant? My completely wild ass guess would be that less-healthy plants absorb more IR light and aren't able to dissipate it, so the thermal is looking for leaves that are holding more heat? I have no idea if that's true or not, I just thought the appeal of thermal for ag was for mostly pre-planting operations. Now I'm curious.
If the above were true, what different issues could cause that spongy layer collapse, and how quickly do they happen? |
|
| |
|
| Thermal is tough for mapping, and gets expensive much quicker.
There's a very select few people that I fully trust as great sources of expertise and information on remote sensing, and none of them are working on thermal right now. That's not to say that you're not 100% correct on the viability, I've just never had any kind of in depth conversation with them on it.
That might be the next topic now, though.
NIR/IR is used fairly commonly right now (and before drones, I mean). How does thermal rank in there? How prevalent has it been over the last 10 years? |
|
| |
|
north east IN | From what I understand as the plant gets sick it will lose its ability to cool itself so it starts to run a fever. Or it will show a higher temp from the hot air then the healthy corn.
You can also use thermal to find field tile as the dry ground over the tile will warm up faster then wet soil |
|
| |
|
| I've never even heard of field tile and had to go look it up. I'm guessing it's not needed as much in areas with permeable soil? |
|
| |
|
southern MN | I understand your are not in production ag, but - wow how different things are in different areas! :)
Half our farmland here wouldn't be farmland without tile, and the rest would be poorer than it is.
While it is obvious as we work through the fields where the wet areas are, being able to identify where existing tile is, and where more is needed is a useful thing around here.
Paul |
|
| |
|
| Thanks Paul, I'll ask around and see if it's used anywhere around here. |
|
| |
|
Eastern Shore of MD | I'm surprised Old&Ugly hasn't chimed in here yet. He has a Flir Thermal camera I believe on an Inspire. I was kind of repeating what I've 'learned' thus far. I can see how it could work in theory like 260david437 said. Heat is often an early indicator of stress.
Like was mentioned, it will be difficult to fly fields every week to get this information. Especially if we are trying to spot very early indicators that can appear and develop rapidly.
I'm not completely sold on the fact that farmers will abandon UAS platforms for hiring a bulk collection service. It all depends on getting the price and performance level right. I personally would rather have my own platform as I would only be flying one crop (250-300 acres) at a time. BUT it has to be affordable and accurate. As far as using the information, I am very interested in trying to import some of these shapefile images into AgStudio for the 'possibility' of actually making fert recommendations based on them and potentially other factors as well.
Edited by BacNBlak 4/28/2016 19:22
|
|
| |
|
| That last part (the shapefiles) is pretty easy to do right now, the problem is just making sure that you're getting GOOD recommendations.
I have a Flir Tau2 turret-mounted on a fixed wing, but I've never used it for any agriculture projects. I might pick up the Vue pro, but I'd like to find someone already using it that can share a few data sets first. |
|
| |
|
Lincoln, NE | Here is the basis behind thermal (see illustration below).
Doing some really interesting things right now with the EMS (400 nm - 13,500 nm) of which I'm not at liberty to talk about.
Personally, I feel what we are working on, will become the sUAS "gold-standard" for Ag remote sensing, but still have a lot of work to do.
To get a sense of the information beyond our visual spectrum - see this video:
https://www.youtube.com/watch?v=kfS5Qn0wn2o
Edited by Old&Ugly 4/29/2016 07:24
 (Thermal EMS.jpg)
Attachments
----------------
 Thermal EMS.jpg (167KB - 729 downloads)
|
|
| |
|
| Thanks O&U. I'm familiar with the info in the video, but your pictures is the start of an answer, at least. ;)
Do you know if there are some published white papers that give more specific information on the function of LWIW in this manner, and the ability to avoid environmental factors? I'm sorting through some old info I've been collecting (trying to find the links that I mentioned in an earlier post), and I found this-
" Yes. Using a uncooled microbolometer camera such as a Flir Tau 2 or DRS core is very dangerous for this line of work according to NASA. There are SEVERAL issues that affect the quality and accuracy of the data, including calibration, atmospheric interference, and the most severe is the presence of any amount of wind that can cause certain areas of the field to suffer from adaptive cooling. Essentially, selling farmers imagery from these cameras is approaching "snake oil".
Also, it is not possible to create an therma orthomosaic without losing radiometric temp data.
The leading person on this matter is: http://aggieair.usu.edu/austin "
I don't necessarily agree with the "snake oil" part, but it all depends in the manner the flights are conducted. It sounds like the capabilities are proven and obvious in a controlled environment, but the environmental variables are the problem (same as what we do, really). Do you have a link to some more indepth reading regarding LWIR and Agriculture? Realistically, I'm not going to be the person developing any kind of software for this, but most of the opportunities for bad/misinformation being introduced happens during acquisition. I just like to be aware of those factors.
Unrelated to the thermal discussion, but here are some links to articles that I was talking about earlier. I'm still trying to find the one that came out of Pendleton, i only have PDFs right now.
http://www.ipni.net/publication/bettercrops.nsf/0/FFCA6C5FBE68758085257D38006CBB7A/$FILE/BC%202014-3%20p7.pdf
http://www.mdpi.com/1424-8220/8/11/7300/htm
http://www.dii.unisi.it/~garzelli/documenti/03-SOME%20DEFINITIONS%20IN%20RADIOMETRY.pdf
|
|
| |
|
| Here is a link to a dissertation on NIR & LWIR use on UAS, if anyone is interested.
http://digitalcommons.usu.edu/etd/2192/ |
|
| |
Introduction from "the other side of the Aisle"
