Editor’s Note: Hemendra Mathur is an investor, board member and mentor to Indian agtech startups. He is a venture partner with Bharat Innovations Fund, a $150 million early-stage fund with a focus on agtech, cleantech, health-tech and enterprise-tech ventures. Mathur previously worked at SEAF India Investment Advisors and Yes Bank. The views expressed here are his own.
In the last few months, we have seen unprecedented coverage of farm-related issues in both print and electronic media in India. Farmer protests in cities like Mumbai and ongoing demand for farm loan waivers across states are receiving wide coverage and with good reason. Risk and inconsistency at various points in the upstream supply chain mean unstable incomes for farmers, especially in a system dominated by smallholders.
Prime Minister Narendra Modi’s target of doubling farm income by 2022 and policy for remunerative prices — set as 1.5 times the cost of production — steps in the right direction. Policy reforms have a huge role to play in making Indian agriculture productive and farm economics rewarding to the farmer. However, setting a policy target of “assured income” for the farmer is probably the right stepping stone in the direction of “doubling farm income”. There is a clear case for prioritizing crop insurance among various farmer welfare schemes and promoting technologies that can derisk farming as well as improve productivity.
A long-term, predictable policy framework can go a long way to address farmer concerns. However, policy reforms take time to build consensus and demonstrate impact, more so in a complex and sensitive sector like agriculture, which is the main source of income for half of India’s population.
So, how can we address farmers’ problems in another way? One answer lies in supply chain innovation and the adoption of technology.
While working with many Indian agtech image-processing startups over the last couple of years, I have realized that image processing integrated with data science techniques can lead to real-time insights crucial for decision making. The images can be captured with various sources such as spectrometer, smartphones, drones, and satellites.
Image processing could be truly disruptive in Indian agriculture in solving inherent problems of productivity, grading and sorting, yield estimation, pest detection, traceability, and detecting food adulteration.
Food Constitution and Contamination
India’s complicated food supply chain, with multiple handling points between farmer and consumer, make it vulnerable to adulteration. Two-thirds of Indian milk gets contaminated on the way to the consumer. Traces of pesticides in leafy vegetables are common and so is a high incidence of arsenic and lead in drinking water. Though compliance with food safety standards is improving with the implementation of a new food safety act, food adulteration continues to be a major health concern.
Portable or hand-held spectrometers are potent tools to detect the type and source of food contamination. The portability and increasing affordability of spectrometers make a strong case for the use of this technology to detect contamination. For example, in dairy supply chains, spectrometers can be used at the village-level milk collection centers, bulk coolers, chillers, processing plant and distribution trucks through to the vending point to detect source, type, and level of contamination.
The imagery from a spectrometer also has the potential to reveal food composition (fat, protein, vitamins, minerals etc.). As new generations of spectrometers get smaller, they may also become an extension of mobile phones. An integrated device, as and when it happens, will be a potent tool in the hands of consumers to conveniently and accurately measure diet as well as check for food contamination.
Scio has developed a pocket-sized spectrometer for food and feed analysis for the consumer as well as business applications. Closer to home, there are startups like Agnext (incubated at IIT Kharagpur), and Testright (incubated at IIT Delhi) which are working on developing solutions in the agriculture and food supply chain, using spectrometers. The challenge is to improve affordability and develop an India-specific digital library for better accuracy and multiple applications.
Making the Grade
One of the key challenges in the Indian supply chain is a lack of defined grades for commodities which prohibits transparent, efficient, and remote online transactions. Mobile imagery can be used for grading on physical parameters such as length, weight, volume, defects etc. For example, Agricx has developed an application for grading of potatoes. Likewise, Intellolabs is working on providing grading solutions for many commodity crops.
Thanks to increasing smartphone penetration and almost universal availability of data (including in remote, rural areas), it is possible to take and transfer images anywhere from farms to local market yards, to distribution centers, to retailers, to consumers. This coupled with blockchain integration can become a strong tool for post-harvest management including monitoring losses (currently estimated at about $13 billion) in the Indian food supply chain Standardization of grades is also key for implementing a pan-India platform (such as eNAM) for the trading of commodities.
Crop and Soil Health
Mobile imagery is also being used for the detection of pests in the crop to suggest pesticides accordingly. There are many startups like Plantix who claim to detect more than 240 pests and plant diseases. Indian startups including Agrostar, Arnetta, Bighaat, and Gramophone are also working on developing image recognition platforms. Many startups are working on soil scans to measure soil nutrition (NPK, micro-nutrients), soil moisture and pH values.
Likewise, drones are being increasingly used for crop health measurement and detecting localized problems such as flooding of agricultural fields, animal intrusion, canal leakage etc. Given the investment and skill set required to fly drones, India needs many skilled “village-level-droneprenurs” who can capture and transfer data from drone imagery, which can be used for analytics and providing timely solutions.
Crop Detection and Yield Estimation
Indian banks are required to lend 18% of their book size to agriculture. However, the majority of banks fail to meet the target for two reasons: lack of adequate data needed for lending and lack of access to millions of farmers. The former problem of lack of access to data can potentially be solved by satellite imagery, which can not only estimate the land area but also detect the crop with reasonable accuracy, which then can be linked to a particular farmer. Many startups like Satsure, Cropin, Agnext, Transerve, Harvesting, Aapah Innovations, Vegamx are working on the use of satellite imagery for such applications. Improving access to free and affordable satellite data (from sources such as Landsat, Planet Labs, and Sentinel) will help in developing more applications around satellite imagery.
There is a significant amount of work going into building algorithms to estimate crop yield by some of these startups. The accuracy levels are improving, and once they start inching towards 90%, they will become important tools to enable crop insurance (in the estimation of farmer risk profile as well as claim settlement).
In a nutshell, a combined approach of using image processing with data science and agronomic tools can address multiple problems facing the Indian food system. The next decade is going to be very exciting for Indian agriculture where image processing is likely to penetrate deeper into supply chains, making them more efficient and transparent.