3D Robotics

Drones vs. crop dusters to gather crop data

water stress

From Steve Blank, the creator of the "Lean Startup" method, in the Harvard Business Review. Key sentence: "They realized that they could use these crop dusting planes to mount their hyperspectral cameras on. They didn’t need drones at all."

Teams that build continuous customer discovery into their DNA will become smarter than their investors, and build more successful companies.

Awhile back I wrote about Ashwin, one of my ex-students who wanted to raise a seed round to build Unmanned Aerial Vehicles (drones) with a hyper-spectral camera and fly it over farm fields collecting hyper-spectral images. These images, when processed with his company’s proprietary algorithms, would be able to tell farmers how healthy their plants were, whether there were diseases or bugs, whether there was enough fertilizer, and enough water.

(When computers, GPS and measurement meet farming, the category is called “precision agriculture.”  I see at least one or two startup teams a year in this space.)

Precision agriculture in practice. Image via Steve Blank.

At the time I pointed out to Ashwin that his minimum viable product was actionable data to farmers and not the drone. I suggested that to validate the minimum viable product it would be much cheaper to rent a camera and plane or helicopter, and fly over the farmers field, hand process the data and see if that’s the information farmers would pay for. And that they could do that in a day or two, for a tenth of the money they were looking for.

(Take a quick read of the original post here.)

Fast forward a few months and Ashwin and I had coffee to go over what his company Ceres Imaging had learned. I wondered if he was still in the drone business, and if not, what had become the current minimum viable product.

It was one of those great meetings where all I could do was smile: Ashwin and the Ceres team had learned something that was impossible to know from inside their building.

Crop Dusters

Even though the Ceres Imaging founders initially wanted to build drones, talking to potential customers convinced them that as I predicted, the farmers couldn’t care less how the company acquired the data. But the farmers told them something that they (nor I) had never even considered – crop dusters (or “aerial applicators”) fly over farm fields all the time (to spray pesticides.)

They found that there are ~2,400 of these aerial applicator businesses in the U.S. with ~5,000 planes. Ashwin said their big “aha moment” was when they realized that they could use these crop dusting planes to mount their hyperspectral cameras on. This is a big idea. They didn’t need drones at all.

Local crop dusters meant they could hire existing planes and simply attach their hyperspectral camera to any crop dusting plane. This meant that Ceres didn’t need to build an aerial infrastructure – it already existed. All of sudden what was an additional engineering and development effort now became a small, variable cost. As a bonus it meant the 1,400 aerial applicator companies could be a potential distribution channel partner.

The Ceres Imaging minimum viable product was now an imaging system on a crop-dusting plane generating data for high value Tree Crops. Their proprietary value proposition wasn’t the plane or camera, but the specialized algorithms to accurately monitor water and fertilizer. Brilliant.

I asked Ashwin how they figured all this out.  His reply, “You taught us that there were no facts inside our building.  So we’ve learned to live with our customers. We’re now piloting our application with Tree Farmers in California and working with crop specialists at U.C. Davis.  We think we have a real business.”

E-mail me when people leave their comments –

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

Join diydrones


  • I like where Mr. Lefebvre is going with the idea but then there should also be automated communication to the irrigation system - as the RMax-like helicopter could not take care of irrigation. So for me, it looks more reasonable to make a cheap small drone, that makes a weekly routine check, brings down the data, which is then automatically processed and the results displayed for the farmer to "press puttons" for irrigation or pesticides etc...

  • Seems to me that to gather the same quality of data, flying at higher altitude and higher speeds which would be required to be able to gather data on the entire area in a single pass, you would need a much more expensive, higher resolution camera system?  So maybe not save so much money?

    And while this approach may work in the short term, it certainly isn't the be-all end-all of this technology.  I think the killer app would be a large robot helicopter (like the RMax) which you push the go-button, and it will head out and completely autonomously look for areas that need treatment, and treat them.  All in one-go. Completely automatically, with all processing being done internally. A true worker-drone.

  • Moderator

    John S, presumably a crop duster providing these services could make some passes over the field for data collection before doing whatever else they were going to do.

  • Developer

    The point of the article is that you can start your business better if you can reduce your costs and cross the chasm between spending money and making money. It's not saying that drones don't have an application. It's about in search of that first customer and push towards a profitable business, they avoided extra R&D costs to prove the business and test the solution.

  • Another thing I find weird about this article is simply the practical aspect: crop dusters fly very low and relatively fast, doing sharp pull ups and dime turns at the end of each beat. By contrast, we want to fly our drones comparatively slowly at around 100m and have them on self-levelling gimbals. The images coming back from a crop duster would presumably be heavily blurred and unusable. Or am I missing something?

  • Developer

    In some near future the actual crop dusting will also be done by UAV's. Japan has been doing this for close to 15 years now (Yamaha RMAX).

  • Maybe this is valid for certain farms/crop types, but I live in a grape-growing region and I've never once seen a crop duster working the vineyards. They do crop dust, but its done by ground-based vehicles. These wine farmers need the imaging done by drones, and that's what we're giving them.

  • I have the Lean Startup book, I need to read it... :)

  • Moderator

    We use manned aircraft to lift our tracking repeaters all the time if we know they are flying off in a direction of interest. I would like to have a couple of trucks have antennas on top and do some collecting whilst travelling. Not so much of a problem but in the future storing and displaying all the different types of data in an easy to understand way is going to be far more important than how the data is acquired. To me as I have said before, low cost flight is sorted now its stupidly simple to set things up compared to a few years ago, data crunching is the less sexy important step.

This reply was deleted.