Drones in Agriculture: Hype or Reality
Unquestionably, #drone is one of the biggest tech buzzwords at the moment. Suffice it to recall recent Y Combinator’s applications “most popular tags” analysis (building IoT-connected AI-powered VR drone?) or 2015 Gartner Hype Cycle, where autonomous vehicles category (one including drones) was placed at so-called “peak of inflated expectations”.
Consequently, agriculture is one of the areas that tech community is bullish the most considering drones usage. In addition, hype is fueled by media with variety of publications claiming how drones will transform the industry and lead humanity to the brave new world.
But is there any hard data supporting this hypothesis?
The first thing to look through in order to get familiar with some industry is market research/forecast. Perhaps many of you are familiar with this pic:
According to AUVSI’s “The Economic Impact of Unmanned Aircraft Systems Integration in the United States” report released in early 2013 agriculture UAVs’ applications would dominate the market and “dwarf all others”. This report is quite foundational and has become a proxy for some other market reports. But, actually, the assumtions on which this report was based are at least very questionable if not entirely wrong.
Here is a great read on this topic, in a nutshell:
- AUVSI report is not an objective piece of research due to this organization’s goal is to “promote unmanned systems”
- Japan is not a good proxy for US due to agriculture landscape is very different, particularly field location, average field size, agriculture products and UAV application (pesticide spraying in Japan and remote sensing in US) itself
However, let’s step aside from research papers towards actual numbers, namely adoption of drones in agriculture.
In 2015 Crop Life Magazine conducted a survey among Ag dealers and producers on how they use precision agriculture technologies. And 2015 survey was the first one that included questions on drones.
Results are pretty eye-opening (here is the full survey):
While 16% of US Ag dealers were offering drone-related services in 2015…
Whereas 16% is not so small number actually, you may think, but it really makes sense to compare this number with satellite imagery services, which are used by 51% of dealers (more on this below).
…only 13% of them were generating a profit from it
And this is really important: doing pretty simple math it can be easily calculated (0,16*0,13), that
Just 2% of US Ag dealers are making money on drones
Such demand definitely doesn’t correlate with AUVSI forecast mentioned above.
Thereby, are there any prerequisites that drone services in agriculture would become a sustainable business?
Drones in Ag: Threats and Opportunities
Down below I’d analyze potential threats and opportunities for drone-related businesses in agriculture.
1. Satellite / Aircraft platforms
By virtue of technology breakthroughs (see “Moore’s Law for Space” and Cubesats) and massive venture capital investments in space-related startups have made small satellite-based Earth Observation (EO) one of the key threats to drone imagery.
Speaking of satellite/drone imagery adoption rate (perhaps the most important metric for assessing one or another technology’s perspetives), UAVs’ adoption among both Ag dealers and producers is at the level satellite imagery has been 11 years ago.
Then, if we look at the historical perspective, satellite imagery adoption rate among US Ag dealers and producers correlates really well with new satellites’ launches:
Thus, taking into account that a huge number of Earth Observation satellites (both by commerical companies and government/academia) is going to be launched in coming years, satellite imagery price will definately decline and this would significantly reduce the barriers for farmers to use remote sensing data.
Moreover, there is a competition from aircrafts which are sometimes claimed to be an optimal platform for gathering imagery in terms of spatial resolution and data acquisition costs at the moment (nice read on this here). And there are some startups utilizing imagery form manned aircrafts, such as YC-backed TerrAvion, which declares it collected more data than the whole electric drone industry combined in 2015
2. Unclear AgTech mid-term adoption
In addition to questionable perspective of drones in agriculture in competition with other imaging platforms, it is unclear how farmers will adopt tech startups’ products in mid-term.
According to Kenneth S. Zuckerberg, senior research analyst at Rabobank Food & Agriculture, commodity price downturn will have serious impact on AgTech startups adoption.
Key takeaway from this study is that farmers would hardly invest their resources to innovations (say provide AgTech startups a revenue) due to significantly declined income.
“Adoption of agtech to drive productivity gains will continue to be delayed and this delay increases the risk of down rounds for startups over the next few years”. — Kenneth S. Zuckerberg, Rabobank Food & Agriculture
In another research paper, made by Jonah Kolb, vice president at farm management group Moore & Warner, and Arne Duss, founder and CEO of HighPath Consulting, even more reasons for limited AgTech adoption are outlined, such as:
- Falling farm incomes
USDA forecasted that 2015 farm incomes were down 36% from 2014, making them the lowest farm since 2006.
- Low incentive
Many US farms are owned by their operators, meaning there is little need to deliver market-rate returns to investors, making adoption of yield-enhancing tech slow.
- Risk appetite
With 62% of US farmers nearing retirement age, there is less appetite for systems upgrades.
- Growing season
A single growing season in much of the US reduces adoption opportunities and the number of potential technology iterations each year.
Summing up, there are some unfavorable macro trends for drones in agriculture beyond competition with satellites/aircrafts.
3. BVLOS Restrictions
The abscence of established mechanism for safely managing UAV operations in low-altitude airspace (at or below 500 feet) and, consequently, beyond-visual-line-of-sight (BVLOS) flight restrcitions imposes significant limitations on the efficiency of drones implementation in agriculture.
While some farms only consist of several acres and could be fully surveyed within-visual line of sight (VLOS), many more farms do not fit this description. For these larger farms the importance of being able to conduct BVLOS operations is magnified.
If farmers with large acreage are restricted to VLOS requirements then they would need to fly multiple, potentially redundant missions to cover the necessary ground. Instead of capturing the imagery and collecting the relevant data all at once, these farmers would be forced to expend precious additional resources into stitching together maps and synthesizing data. This would be highly inefficient — both in terms of manpower and time — and could nullify the potential time and cost savings provided by UAS for Ag industry.
At the same time, there are some technical (such as lack of access to spectrum and uncertain UAS traffic management (UTM) system architecture) and regulatory barriers that may cause the system to be years away (more on this here).
Moreover, BVLOS are limited in most of the countries (not only in US):
Thus, BVLOS restrictions is one more threat to potential expansion of UAVs in Ag.
Despite the threats described above, there are some positive signals making drones’ perspectives more promising.
1. Drones Are Getting Smarter & Stronger
Generally, drones implementation is limited due to several technological difficulties, such as:
- Lack of autonomy
UAVs full potential can be unlocked only when truly autonomous drones would be available.
- Low flight endurance
Efficiency of drones operations strongly tied with the flight endurance, which is, for most of the professional UAVs, such as SenseFly eBee and AgEagle, is about 30-40 minutes, which is not enough obviously for continous surveys.
However, a lot of great startups are overcoming these challenges with their products.
Recent computer vision applications improved drone capabilities beyond pretty straightforward “follow-me” features towards impressive autonomy. To name a few startups in this area there are Skydio (raised $28M from Accel and a16z) and Percepto (raised $1M from Mark Cuban and some other high-profile angel investors). Moreover, advanced computer vision is already imbedded in consumer models, such as brand new DJI Phantom 4.
As for flight endurance improvement, there are 2 options: ground charging stations and advanced (not LiPo) batteries.
Regarding batteries, there are several projects developing fuel cells for UAVs, such as hydrogen fuel cell by Intelligent Energy. Moreover, world record for the longest drone flight (more than 3 hours) that was recently set in Russia involved hydrogen-air fuel cells.
Thereby, considering that drones would be capable to fly for hours and do it without humans’ assistantship quite soon, this will significantly increase drones’ operations efficiency.
2. UAVs Sensors’ Advancements
One area in which drones are definitely ahead of satellites at the moment is sensors variability and data resolution. With lidars , hyper-/multispectal imagers and thermal sensors drones are capable to collect unique data compared to satellites. For example, an average hyperspectral image space-based system can provide has 30–50m GSD, which is 2 order of magnitude lower than it’s possible to get using UAV.
But, while these sensors have already proved its value for construction, mining, energy and O&G, it remains unclear which applications (except NDVI calculation utilizing multispectral sensors, which can be successful provided with satellites as well) for agriculture can be useful for farmers (crop counting or weed detection?).
Moreover, recent announcement of a Planetary Recourses Series A round for the development of hyperspectral and thermal imaging satellite constellation and company’s partnership with Bayer signals an increasing competition between drones and satellites in the area of “uncommon sensors” also.
3. Ag Can Adopt New Tech Pretty Fast (if it brings real value)
Despite the data on satellite/drone imagery penetration and unfavorable AgTech mid-term adoption forecast mentioned above, historically, Ag industry have snapped up some technologies pretty fast, such as GPS-related ones.
Thus, if we take these technologies as a reference to UAVs and extrapolate adoption growth rate we’d have pretty optimistic scenario with around 60% (taking GPS guidance with auto control/autosteer 30% CAGR as a reference) penetration of drones in US agriculture industry till 2020.
In my opinion, unfavorable conditions prevail over the favorable ones at the moment and global expansion of agriculture drones is questionable (at best).
However, there is probably some niche on the intersection of:
- Hyperspetral Imagery
UAV-based hyperspectral imagery is much more affordable than the one from satellite-based sensors at the moment (and in a mid-term too, taking into account Planetary Resources constellation will probably be fully operational at 2019–2020). Thereby, if startups would be able to solve some technology problems related to hyperspetral data (huge data volume, the need for calibration for different geographical areas) and create uses cases providing value to the Ag customers — it’s is one of the things that would give drones an edge compared to other platforms (an interesting company in this area: Gamaya)
- Relatively small farms
At the moment, to purchase satellite imagery it is required to buy some minimal order amount of data (e.g. 500 km2 for 5m GSD data from RapidEye constellation), which makes it unreasonable for small farmers. Consequently, small farmers represent a market opportunity in agriculture for drone startups, perhaps Japan and Western Europe, speaking of geography. However, there are some limitations on the area on which drones’ implementation is economically reasonable (here is a research paper on drones, aircrafts and satellite cost effectiveness comparison:
Thus, drone Ag startups probably should try to target a niche market/use case, establish a beach head instead of aiming for global expansion from the beginning and prove they can really bring some value to the farmers (and provide return to its investors consequently).