We Have Adopted All-Inclusive Approach to Address AI in Agriculture Market: VP, Persistent Systems
By integrating Internet of Things and machine learning into the Agriculture sector, the future of Agriculture seems to look brighter, as efficiency and productivity of crop management are poised to grow simultaneously
India is in need of evergreen revolution, said MS Swaminathan in a conversation with BW Businessworld. It is a truth that none can deny, with more than 1.3 billion people to feed on the domestic front, Internet of Things (IoT) may be a fruitful tech help for better integration at various levels of farming. Precision agriculture (PA) is an approach to farm management that uses information technology (IT) to ensure that the crops and soil receive exactly what they need for optimum health and productivity.
By integrating Internet of Things and machine learning into the Agriculture sector, the future of Agriculture seems to look brighter, as efficiency and productivity of crop management are poised to grow simultaneously. BW Businessworld Speaks with R Venkateswaran, Sr. VP IoT of Persistent Systems, one of the major players in the field of Precision Agriculture.
How relevant is it to identify the scope of the application of Artificial Intelligence at a farm-level with a huge gap of technical training on ground level?
It has been our experience that many of the relevant AI technologies work behind-the-scenes and hence the lack of technical training on the ground level is not necessarily an obstacle to adoption. The scope of technology deployment usually spans across all farmers in the entire village or even clusters of villages. To encourage wider adoption, we have focused on simple, intuitive mobile-app/SMS based interactions between the farmers and the backend AI systems; and hiding the complexity from the farmers.
We have been working on some of the challenging AI applications that help the farmers with automated crop-related advisories and recommendations. To address these challenges, we apply machine learning on the data collected from the variety of data sources. Some of the applications of AI/ML (machine learning) include:
i) Recommendation of the appropriate crops specific for the farmer based on soil type, water availability, temperature, humidity, size of the farm.
ii) An early warning system for pest infestations, including the recommendation for the appropriate pesticides with dosage to prevent/minimize the impact of infestation. Prevention of pest infestation requires a complex correlation of factors such as crop type, crop maturity, weather and other environmental conditions. For small farms, pest infestation levels can be detected through a combination of image analysis (leaf colour, partially-eaten leaves etc. and crowd-sourcing (by farmers). The pesticide and dosage recommendation requires subject matter expertise to be in-built into the Machine Learning system.
iii) Recommendation engine to plan harvesting activities that maximize the yield. This recommendation is based on the maturity of the plants inferred through image analytics, prediction of imminent adverse weather, availability of resources such as farming equipment and manpower etc. For example, it may be advisable to harvest prematurely and get a smaller yield if a prediction of adverse weather (snow/rain) is likely to destroy the entire crop.
Given the bandwidth constraints and the remoteness of the farmlands, our technology solution leverages a combination of Edge-based Machine Learning and Cloud-based Machine Learning. This functional split of ML is key to the success of AI/ML-based solutions wherein we get the high responsiveness of Edge ML along with the powerful computational capacity of Cloud ML.
How strong will the competition in India for precision farming?
In the past, the Indian agricultural segment spanning 75% of India’s population across 600,000 villages, which has not been viewed as a very lucrative market for technology adoption. However, with the increased focus on Digital India initiative, the digital divide is narrowing. Today, the availability of affordable sensors and ubiquitous cellular coverage can potentially be a game-changer as they facilitate data collection which is a key necessity for ML/AI based solutions to work.
Precision Agriculture can also benefit immensely from other government initiatives and systems such as Crop SAP that have access to vast quantities of agricultural data. Making this data available to different vendors will encourage them to explore opportunities and build solutions for the farmer community.
Precision farming with a micro-focus on a small group of farmers, combined with AI/ML techniques for predicting adverse events and recommendation for corrective actions, can uplift the lives of the farmers significantly and bring them above the poverty line. In the coming years, this will spur the technology companies to revisit their strategy for addressing this potentially under-served market.
Who is your major partner, government or individual level farmer?
We have adopted an all-inclusive approach in addressing this market. On the technology front, Persistent’s IP under the Concert IoT product suite is well positioned to address the challenges of the farmer community. We partner with hardware vendors for agriculture-related sensors, connectivity vendors for network coverage, software vendors that integrate with the Concert IoT product suite and thus provide end to end solutions.
We work directly with the Government agencies, as well as partner with non-profit organizations working on rural uplift programs. We also leverage some of our CSR initiatives to reach out directly to the specific set of villages. This helps us to build solutions based on the extensive domain knowledge garnered first hand and through our association with such organizations. It is our constant endeavor to ensure that the benefit of the technology solutions is directly reachable to the farmer.
How will ecosystem and monetization enablement help organized farming players?
Farm ecosystems include both the supply chain and distribution chain. Supply chain data monetization can help suppliers produce and deliver customized products (such as fertilizer and pesticides) that are tuned to the unique needs of a specific farm or region. Distribution chain data monetization can help farmers to ensure the better longevity of products sold to the consumer by carefully monitoring and optimizing the harvest, cold storage, and transportation logistics. Finally, ecosystems allow regional farm cooperatives to develop and share precision agriculture knowledge (in the form of algorithms) that leverages regional best practices and learns from the data generated by farmers in a region.