China is facing a serious problem on its farmlands. Contamination from industrial, mining, and agricultural activities is affecting soil quality on the country’s arable land so severely that nearly 20 percent of China’s arable land is no longer farmable, according to a 2014 survey by China’s Ministry of Environmental Protection and Ministry of Land and Resources. “That amounts to roughly 250,000 square-kilometers of contaminated soil, equivalent to the arable farmland of Mexico,” The Economist reported in an article titled “The Bad Earth,” published earlier this year. “Officials say that 35,000 square-kilometers of farmland is so polluted that no agriculture should be allowed on it at all.”

China’s heavy and insufficiently regulated industrial sectors are major contributors to the problem. Indeed, cadmium and arsenic, which are released into the air from smelting plants and other industrial sites, have been found in 40 percent of contaminated soil samples.

The problem did not start with heavy industry, however. Soil samples can tell stories that are decades, even centuries, old. “China’s chemical and fertilizer industries were poorly regulated for decades, and the soil still stores the waste that was dumped on it for so many years,” according to The Economist.

With a population of 1.4 billion people to feed, and just 0.08 hectares of arable land per person—including land degraded by pollution—China cannot afford to lose any more farmland. Today’s farming practices do not offer much hope for soil remediation: “Since 1991 pesticide use has more than doubled, and the country now uses roughly twice as much per hectare as the worldwide average,” the article continues. “Fertilizer use has almost doubled, too.”

Nearly 20 percent of China’s arable land is contaminated. That’s roughly equivalent to Mexico's total arable land.

A significant problem relating to pesticide and chemical fertilizer use in China is the method of application. At least 95 percent of Chinese farmers still spray their farms by hand using a backpack sprayer, a tank of the chemical, a pressurizing device, a pipe, and a metal wand with a spray nozzle at the end. Applying pesticides by hand makes it hard for farmers to regulate the amount of chemicals they use, and some spray their fields as many as 15 times per year.

Beyond the environmental impact, exposure to toxic fumes and contact with chemicals from hand spraying poses considerable health risks to farmers. For example, the Center for Disease Control and Prevention in China’s Zhejiang province registered 1,413 pesticide poisoning-related deaths among farmers between 2006 and 2010. Non-fatal poisoning cases during that period topped 20,000 during that period.

“In today’s world, these deaths are avoidable,” says Jan Gasparic, head of enterprise marketing at Chinese drone manufacturer DJI. “That’s what our technology is setting out to do.”

In 2015, DJI introduced the Agras MG-1, its first crop-spraying drone, as an alternative to the dangerous methods of pesticide application Chinese farmers and their workers employ.

Agtech revolution

For comparative purposes, in the U.S., pesticide application is different. American farmers typically apply crop protection chemicals using boom sprayers, which are spray nozzles attached to tractors that can spray several rows of crops at a time while the farmers are a safe distance away, in the cab of the tractor.

There are several reasons why such tools are not readily used in China. For one, the average farm in China is only 0.65 hectares in size. China’s 260 million rural households work 120 million hectares of farmland, according to the Research Centre for Food and Agricultural Economics at Nanjing Agricultural University. They occupy such small plots of land as a legacy of Mao Zedong’s Great Leap Forward, which nationalized China’s farmland then later divided it into tiny plots to enable families to feed themselves following a major famine that killed as many as 30 million people between 1959 and 1961. The northern part of the country does contain larger, government-run farms of a few hundred hectares each. Overall fragmentation of the agricultural sector, however, makes China’s crop production inefficient and costly.

There are geographical challenges to deploying larger-scale machinery too, such as uneven terrain. China’s rice terraces are layered shelves of land that would make it difficult if not impossible for farmers to use rolling tractors with spray booms. Chinese farmers therefore undertake or hire cheap labor to do many of the tasks that machines perform on U.S. farms.

That is slowly changing, however; the technological revolution that is hitting the global agriculture sector is also beginning to impact China. Drones in particular have been making waves in numerous agricultural markets. Also called unmanned aerial vehicles (UAVs) or unmanned aerial systems (UAS), drones are flown by ground-based control systems that are piloted by a human operator on the ground or autonomously with an onboard computer and piloting software. Drone technology was first employed by militaries for missions too dangerous or difficult for humans. The technology was later scaled down and adopted for consumer and commercial use. In the U.S., the promise of drone technology in revolutionizing farming practices reached a fever pitch in 2015, when an onslaught of new hardware, software, and services offering everything from crop monitoring to field imaging hit the market.

The promise of drone technology in revolutionizing farming practices reached a fever pitch in 2015, when an onslaught of new hardware, software, and services hit the market.

Shenzhen-based DJI made a reputation for itself when it launched its “Phantom” quadcopter imaging drone. The product has secured the 11-year-old company a US$10 billion valuation and its place at the helm of the consumer drone market. The company now produces multiple drone series for different purposes, but its most popular UAVs continue to be its imagery models, which have been adopted by American farmers in addition to hobbyists, filmmakers, and other professionals. With many U.S. farms stretching several hundreds hectares, it is nearly impossible for farmers to keep track of what is happening on every meter of their land at all times. Farmers have deployed the camera-equipped Phantom drones to survey fields and collect information on how crops are growing. Farmers can then analyze the images to decide how and when to intervene if crop health appears compromised.

Drones for imagery have not taken off in China’s agriculture industry the same way. Thus, when DJI launched its Agras MG-1 model in November 2015, it was designed specifically to tackle the negative impacts of hand pesticide spraying in rice terraces of Southern China. DJI claims that the Agras MG-1, and its successor the MG-1S, are 40- to 60-times faster than manual spraying operations and can cover 4,000 to 6,000 square meters in 10 minutes. That equates to 65 to 80 acres a day.

How does it work?

The Agras MG-1 is an octocopter, meaning it has eight propellers. At 1,471mm by 1,471mm by 482mm, it is on the larger size of most of DJI’s drones, which typically hold just a small camera. Without its propellers, however, the model can fold down to half its width.

The MG-1 has four spray nozzles, each placed directly below a motor, and is built to carry up to 10kg of liquid payload. The spray nozzles are interchangeable to optimize atomization, energy efficiency, and the liquid spray rate for different types of chemicals. The downward airflow generated by the propellers accelerates the spray, and helps to maintain a constant flow, even amid changing flight speeds.

DJI’s flight controller software is programmed for three flight modes: “smart” mode, “manual-plus” mode, and manual mode. With the smart mode, pilots can map out flights with the controller, including specific chemical application parameters, which are then undertaken autonomously using real-time kinematic satellite navigation. The MG-1 automatically records its coordinates as it progresses across a field, so that the drone can resume its flight plan from the last point in memory if it needs to pause for a chemical refill or for battery charging. Pilots can also save specific flight plans to use at a future date.

In manual-plus mode, the MG-1 will make the same turns as in the smart mode at the click of a button. Manual mode, meanwhile, is for pilots who want to freely control the drone in real-time.

Of the three modes, smart mode is particularly helpful for pesticide application, because it can be hard for manual operators to maintain a consistent flight height over uneven fields, which are common in many parts of China. That can affect the spraying accuracy and thus the efficacy of the crop input. In smart mode, however, the MG-1 is able to constantly adjust its height above the crops according to the terrain, because it has built-in sensors on the front slope, rear slope, and bottom of the spray tank, which can scan and adjust the drone’s altitude to match the terrain in real-time.

Gasparic says that the combination of the above technological features ensures that the MG-1 sprays within seven millimeters of precision.

This year, DJI launched an upgraded version of the MG-1: the Agras MG-1S. “We invested a lot in flight planning and carried some features over from our other drones to the MG-1S to improve it,” says Gasparic. With the newer version, a pilot can chart the boundaries of a field with the controller, and the drone will create flight routes accordingly. Also, whereas the MG-1’s controller does not include an integrated tablet, the newer model’s controller has a bright 14cm, 1080-pixel display to make the field mapping process easier. “That’s really helpful in laying down the flight plan, especially if you need to exclude certain areas,” Gasparic adds.

Area exclusion and route automation are vital for agricultural drone operators in northern China, where farms are larger. “All that intelligence is built into the drone so that the operator can do less work manually. In the South, a drone pilot flying the vehicle manually [is] sufficient,” Gasparic explains.

The MG-1S also has a new, intelligent spraying system that increases the accuracy and precision of the application. With two compatible pumps controlling the front and rear nozzles separately, the device can operate three spraying modes: forward spraying, backward spraying, and full spraying. A new pressure sensor and flow sensor monitor the spraying rate in real-time, enabling the drone to adjust the spray volume based on flying speed. The MG-1S also allows pilots to regulate and record how much pesticide or fertilizer is applied—a feature that has both cost and environmental benefits.

Practically, both models of the Agras drone are robust hardware made from high-strength carbon fiber materials, which make the frame lightweight but durable enough withstand harsh conditions. Both versions are collapsible, making them easy to transport, and can be easily dismantled for repair.

Also, because equipment used in plant protection operations is susceptible to dust and corrosion, the Agras models are designed to avoid corrosion and degradation with a sealed body and an integrated centrifugal cooling system. As it flies, air enters the aircraft’s body via a front inlet, which is equipped with a triple filter system that keeps dust and other particles from getting inside. The air then passes through each of the aircraft’s arms to the motors, capturing heat from the internal components and dispersing it into the surrounding air. These cooling and filtering features increase the motors’ expected lifespan threefold, according to DJI.

Different markets, different drones

Gasparic likens the adoption of crop spraying drones in China to the adoption of mobile phones in Africa, which leapfrogged land-based telephone technology. “In China, there was a specific opportunity for us around using drones as a spraying technology, based on the existing market conditions,” he says.

By contrast, crop spraying drones have yet to take off in DJI’s biggest market, North America, because of the drones’ limited spraying capacity relative to the size of most North American farms. Having the ability to spray only 65 to 80 acres per day is inefficient on farms that are hundreds of hectares in size. That trend could change when drones with larger payloads and longer battery life become available. In the meantime, imaging drones continue to be the dominant UAV technology in the U.S. agriculture sector.

Gasparic is not necessarily surprised that the two markets have embraced two entirely different drone technologies. “It’s contextual, meaning some technology might not necessarily fit in the same way and place in China [as it does in the U.S.],” Gasparic says. “Using a drone for spraying chemicals is solving a very different problem to remote sensing.”

“Sensing shows farmers areas that might have problems, and even tries to identify those problems, enabling farmers to quickly [respond] with water, or pesticides, or nutrients,” he continues. “Spraying is directly administering [an] input, so those two platforms serve different purposes completely.”

Another reason Chinese farmers may not have embraced imaging drones the way U.S. farmers have is because there is less of a culture of integrating data and software into farming operations in China. “Our remote sensing drones get a lot of interest when we take them to all of our agriculture shows, but [Chinese farmers] are still getting acclimatized to the idea of taking data and uploading it to the cloud. People are not as comfortable [with that] in China as in the U.S.,” Gasparic observes.

This is in part because the Chinese agriculture sector does not have the same structure of advisory networks of agronomists and suppliers that the U.S. sector has. “For example, variable rate applications are accepted and understood in the U.S. as an important way to reduce the amount of inputs on the field,” he explains. “But in China, the chemical companies are working on a different scale, with lots of sponsorship from the government, so the economics and drivers for adopting that technology are different. It’s almost the polar opposite.”

Adoption dynamics

The Agras models have sprayed 1.3 million hectares of farmland in China since launching in 2015. That is still less than one percent of the 515 million hectares of agricultural land in China.

DJI recognizes that it is still in the early days of introducing its technologies into China’s agriculture sector. “You can’t just parachute a technology into an industry and expect adoption to go,” says Kevin On, DJI’s director of communications. “It needs support though training and education and nuanced positioning [catering to] the market needs.”

One of DJI’s strategies for encouraging uptake of its Agras drones is partnering with Chinese spray service companies. Many farmers in China employ contractors to undertake certain farming because it is cheaper and more efficient than doing those tasks themselves. Crop spraying is one of those. Hiring contractors also reduces the need for the farmers to purchase equipment themselves.

For crop spraying service providers who are used to manual spraying, DJI’s drones are good for their businesses, says Gasparic. The MG-1’s digital management platform is particularly useful for keeping track of multiple aircraft and missions. The drone’s platform can supervise the flight status of every aircraft in a service provider’s fleet, as well as assign fields to individual operators and check on completed fields.

DJI estimates that the MG-1 can reduce service providers’ costs by 30 percent, while giving farmers a better level of service at a similar price point to what they would pay for a hand-spraying service. “Before, with hand sprayers, these providers were sometimes spraying too much or too little, so it was inconsistent,” says Gasparic. The service providers can show evidence of their field coverage to farmers via the MG-1’s droplet camera.

“You can’t just parachute a technology into an industry. It needs support though training and education and nuanced market positioning.”

Another approach DJI has taken is launching the Unmanned Aerial Systems Training Center (UTC) to train drone pilots. The UTC’s current programs range from general operation and maintenance to specialized courses in aerial infrared measurements, advanced aerial photography, and UAV software development. The center also has a course specifically geared towards agriculture, for which pilots can receive certification after undergoing the training.

“UTC has helped to set the standard for how our technology is deployed in agriculture,” On says. Trainees in the agriculture course include private drone owners, drone providers, and employees of spraying businesses.

DJI has trained 1,000 pilots through the UTC so far, and the company is hopeful that the program will bring new talent into the agriculture industry. Already, some of its trainees in the UTC have included graduate students and young entrepreneurs setting up done-based businesses that provide services to farmers in their local communities.

“It is really interesting to [DJI] to be creating jobs and encouraging startup businesses to start their own services for farms,” says Gasparic. “We’re seeing an ecosystem build up around what we’re doing in China.”

This will be increasingly important as more and more of China’s would-be farmers flock to cities for work opportunities instead. A recent article from Bloomberg reported on a year-old company called Hainan China Agriculture and Flight Service, which is offering spraying services for farmers who have had difficulty finding workers to spray their fields by hand. “It’s harder to find people to spray pesticides in the old way. Young people want to leave the farms and find better jobs in cities now,” the article quoted one farmer as saying.

DJI is beginning to explore the idea of entering new markets where farmers are expressing an interest in UAV technologies. But the company is proceeding cautiously.

“There are a lot of service providers enquiring about the platform and how to use it in their home markets, for instance in Japan, Malaysia, Korea and Indonesia where plantations and farms are a big part of GDP,” says On. “But first we need to understand how applicable the technology will be to them.”