Eric Marx reports on apps such as Plantix and the Sowing App, which are helping smallholders achieve optimal harvests in some of the harshest growing areas on the planet

As digitisation rolls across India, some see the dawning of a second green revolution in the agricultural sector. Whether, how, and with what success it will affect the lives of India's poorest and most marginal farmers – so-called smallholders, who typically work just over a hectare of land and account for roughly 65% of India’s population – is still an open question.

“India is a very interesting place to be working on digital agriculture because there are so many agri-tech companies getting involved,” says Anthony Whitbread, a research programme director of drylands innovation systems at the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), an international non-profit based in India that has been experimenting with the latest in digital technologies in dryland regions across south Asia and sub-Saharan Africa.

India is a very interesting place to be working on digital agriculture

Up to now there have been few examples of large-scale impact that have been immediately useful for smallholders.

One is the Plantix App, which uses image recognition technology and self-learning algorithms to enable real-time monitoring of pests and diseases. Another is the Sowing App, which helps farmers achieve optimal harvests by advising on the best time to sow using data about local weather and soil conditions.

Both were developed by ICRISAT, working with private agri-tech partners adept at leveraging the latest advances in big data and artificial intelligence (AI).

PEAT says there has been a surge in farmers sharing knowledge. (Credit: ICRISAT)

Mobile phones and cheap data plans linked to better analytical capacity are helping to drive this new approach to farming, called precision agriculture, says Whitbread. So are improved satellite data (including better resolution, increased frequency rates and lower costs) and cheaper remote collecting devices (including sensors, global positioning systems and drones), powered by the internet of things.

Together, this suite of technologies may hold a key to solving one of humankind's biggest challenges in the 21st century: ensuring food and nutritional security to a world population that is estimated to swell to 9.3 billion by 2050.

India alone will have to feed 1.4 billion people by 2025, against a background of degrading land and increasingly variable weather associated with climate change.

The key is enabling famers to get the right advice for free in real-time, and then allowing them to connect with other farmers

But given that most smallholders are not digitally literate, typically work with their bare hands in fields lacking irrigation, and have limited access to markets due to poor roads, is it reasonable to think that digital technology can markedly change their prospects?

Whitbread concedes that there is a lot of “hype” around digital agriculture, but adds that there is new evidence of ways to make it more accessible to small farmers.

First, the technology must be simple. Second, governments should work in partnership with the private sector, academic institutions and non-profit research and development groups to keep costs affordable. The key is enabling famers to get the right advice for free in real-time – and then allowing them to connect with other farmers so they have access to a range of different markets and services.

The Sowing and Plantix apps show the way, though by themselves they will not revolutionise farming, Whitbread cautions,. For that to happen there will have to be a third development: progress in logistical and market-type digital applications, as well as digitised crop insurance.

Public agricultural research institutes like ICRISAT know very well that Big Ag companies are going to come into the picture. In fact, the seed traders and fertilizer salesmen are already there, says Srikanth Rupavatharam, a digital agriculture scientist with ICRISAT who coordinates a network of “photo-hunters”, which are now tagging and uploading images with the Plantix App.

What Plantix does is re-balance that relationship so farmers and government-sent extension workers, whose mission is to educate farmers in the latest in scientific research and agricultural practices, get independent and accurate information that they can use in a timely manner.

A rice farmer uses the app to check for disease. (Credit: ICRISAT)

Plantix’s core feature is image-recognition software, which uses self-learning algorithms to detect and recognise optical patterns in photographs of plant damage caused by pests, diseases or nutrient deficiencies, explains Bianca Kummer, a co-founder of Progressive Environmental and Agricultural Technologies (PEAT), a German-based company started by young entrepreneurs who first tested the Plantix concept on greenhouse tomato plants three years ago at the University of Hanover.

“At the beginning there was no funding money, only an idea and seven highly motivated people,” says Kummer.

The partnership with ICRISAT has been essential, given the institute's deep network, including scientists at state agricultural universities and extension workers at the department of agriculture in Andhra Pradesh, a province in south-east India where testing with smallholders is now taking place.

You have to train the network to actually use AI to identify these pests

“This is the core thing – getting databases for each pest and disease – because you have this AI but you have to train the network to actually use AI to identify these pests,” said Rupavatharam.

The more data one has, the more complex the probability calculations and, in turn, the greater the accuracy. Plantix currently has a database of more than three million pictures. But for every pest or disease to be approved, Rupavatharam says, 1,000 to 2,000 pictures have to be scientifically tagged.

The app can detect upwards of 340 diseases covering 20 crop types, including rice, wheat, millet, maize, potato, cucumber, eggplant, lentils and ground nuts. For 120 of those diseases the accuracy rates are already as high as 95%, says Kummer.

Monsoon season; the link to weather forecasting is crucial. (Credit: Daniel J. Rao/Shutterstock)

PEAT will soon expand the service to include a Plantix crop calendar, which takes into account specific parameters given by the farmer, merged with regional data like soil type and weather reporting, alongside a community forum feature so that, in Kummer's words, “we can accompany the farmer throughout the year starting with seedbed preparation and go all the way to harvesting time.

“Once we have created a bigger reach out, this ‘ground truth data’ will be offered in a processed way to all stakeholders in the agricultural sector who are interested in tracking diseases in their spatial and temporal expansion.”

Possible customers for this data could be micro insurance companies, governments and non-governmental institutions, as well as input companies for seeds and fertilisers.

The crop calendar is important in understanding the benefit of trying to treat a disease or pest

In a second step, adds Kummer, she and her team can imagine the creation of a digital marketplace that would charge the seller who uses Plantix for contacting the buyer. For the farmer, the app would always remain free, with the diagnostics devoid of any company brand names.

The link to weather forecasting is especially important. “When the crop is close to maturity much of the time you don’t spray or try to control a disease, says Whitbread. “So the crop calendar is important in understanding partly the economic benefit or practicality of trying to treat a disease or pest.”

A farm in Anantapur district, where the app is being trialled. (Credit: Macsimilion/Shutterstock)

But that also brings in more complexity, and a big part of the reason for the Plantix app’s quick scale-up has been its simplicity.

The Sowing App, also referred to as the intelligent agricultural systems advisory tool (ISAT), takes the complexity one step further by modeling historical and actual climate data together with local soil type and rainfall data.

Microsoft helped ICRISAT code the information into an online system that sends push notifications to the farmer. Before the season the farmer gets a message about the likelihood of a good season, plus the optimum sowing times. Once the conditions are exactly right, an additional message is sent, followed by weekly messages about management, cultivation and inputs they should be considering.

We think if we can get it right in this area, we can scale it up anywhere

“And the idea now – which Microsoft has backed pretty hard – is to scale the project by automating these processes with the use of sensors,” says Whitbread.

To this point ICRISAT has limited the project to one crop, groundnut seeds, while testing it on 4,000 farmers from Anantapur, a semi-arid district in Andhra Pradesh that presents some of the harshest conditions for growing ground crops.

“We think if we can get it right in this area, we can scale it up anywhere,” Whitbread says.

High-value crops, such as tea, help attract funding. (Credit: T photography/Shutterstock)

Going forward, ICRISAT and other development groups want to scale up the effort to include more technology and public-private partnerships. According to the institute, the goal is to fund start-up ecosystems in order to launch business models that exist outside of donor and foundation funding. In that way, ICRISAT and its other development partners can head off “private capital [that] may begin to define products and approaches that exclude poorer farmers”.

Central to that effort is ICRISAT's iHub incubator, which opened in February 2017 and is currently home to 10 companies working on everything from cloud-based seed monitoring and the launch of low-cost weather stations to the production of drones for agricultural operations.

Everybody thinks, ‘ah, some lovely new technology, that’s going to feed the world.’ It's hugely hyperbolic

Camilla Toulmin, an economist with the International Institute for Environment and Development (IIED), welcomes the new focus on including start-ups that could help agricultural production and give farmers greater voice, but has reservations about the propensity ICRISAT and its sister institutes have had towards collaborating with “very large companies and not the many small and medium enterprises (SMEs) which might benefit from their knowledge”.

Moreover, what may work in India – where there are highly networked, very tech-savvy people operating in a broadly supportive government environment – may not work so well in other places where there are insecure tenure rights and greater political strife, she says.

“Everybody always thinks, ‘ah, some lovely new technology, that’s going to be brilliant. That’s going to feed the world. That’s going to save poor farmers from poverty.’

Camilla Toulmin warns against hyperbole around new technology. (Credit: IIED)

“It's hugely hyperbolic in terms of the promises made, and in terms of what subsequently happens,” says Toulmin. “So I would be a bit cautious. It will take more than new tech to ensure they have the access to the land, water and political power they need to make a real difference to their lives.”

Jerome Bossuet, a spokesperson at ICRISAT in its partnerships division, responds to the criticism by saying that the private sector can clearly see the advantage of having a safe and efficient supply chain for high-value crops such as tea, coffee and tobacco. For other crops, he says, “such systems may help improve farmers organisations’ linkages to the marketplace,” meaning the benefits are going to them directly rather than to intermediaries.

It is not yet a magic answer to everything, but at least it’s helping farmers to have a voice

Rupavatharam says at the very least, mobile apps such as Plantix are highly empowering. He says preliminary results from a survey now being conducted by PEAT indicate an upsurge in farmers sharing knowledge about traditional methods once a diagnosis is received from the app. Upwards of 25,000 messages have been posted in a single day.

“I can't read them all!” exclaims Rupavatharam. “That is one of the most important things that is happening here: it is not yet a magic answer to everything that is happening with plant protection, but at least it’s helping farmers to have a voice. That’s where I see the strength of this app.”

Eric Marx is a Berlin-based journalist covering issues ranging from green energy to sustainable sourcing, climate change and transparency. He has served as a correspondent for ClimateWire News and as an energy reporter with Montel newswire


This is part of our in-depth briefing Future of Farming. See also:

Future Farming: Pastures new for big data

From vertical farms to new proteins: innovating to feed the planet

A sea change in the way we farm fish

Applying AI to the humble potato

Main picture credit:


PEAT  climate-smart agriculture  India  SDG2  smallholder farmers  agri-tech  ICRISAT  Plantix  IIED  Sowing app  precision agriculture  Microsoft 

comments powered by Disqus