By Anthony Devassy
With depleting resources, reducing land sizes and increase in input and labour costs, combined with uncertainties around weather and market prices, agriculture in India has become a profession that is more risk intensive. This makes crop insurance a very critical vehicle for mitigation.
The farming window is very restricted and delays in claim payments can often hinder a farmer’s prospects for the next farming season. Often, insurance companies tend to dispute yield loss data sent by states. The Pradhan Mantri Fasal Bima Yojana (PMFBY), a government sponsored crop insurance scheme, was introduced to integrate multiple stakeholders on a single platform. Unfortunately, many processes related to the government schemes are manual, leading to delays in claim payments.
This is where Artificial Intelligence (AI) and Machine Learning (ML) can help augment the processes related to the current crop insurance schemes, both in preventing the claim delay and in reducing the timeline in claim settlements.
Monitoring crop growth using IoT platform/GDD Model: As soil temperature influences crop emergence, ambient temperature is a crucial factor for crop development. Growing Degree Days (GDD) model uses hyperlocal temperature, humidity to estimate the growth and development of plants during the growing season. Field deployed IoT sensors data such as ambient temperature, soil moisture, relative humidity can be ingested real-time into an AI/ML platform to perform spatiotemporal analysis and develop a GDD model to predict estimated crop growth and yield at farm level.
Weed control: Computer vision could aid in precision spraying which can help bring down the amount of chemicals that are sprayed on crops and reduce the herbicide expenditures.
Soil health monitoring: Though the government came up with a Soil Health Card (SHC) scheme, Indian farmers have not been benefitting. Deep Learning solutions could help identify potential defects and nutrient deficiencies in soil. Complex algorithms help correlate foliage patterns with certain soil defects, plant pests and diseases. The image recognition capability could help identify possible defects through images captured by a user’s smartphone camera.
Drones and computer vision for crop analysis: Drones in agriculture could be pursued for soil and field analysis, planting, crop spraying, crop monitoring, irrigation, and health assessment. The device will leverage computer vision to record images which will be used for analysis. Algorithms are used to integrate and analyse the captured images and a detailed report on the health of the crop can be provided.
Satellites for weather prediction and crop sustainability: ML could be deployed in connection with satellites to predict weather, analyse crop sustainability and evaluate presence of diseases/ pests.
Imagery for crop cutting experiments (CCE’s): Crop insurance norms require four crop cutting experiments at every village which translates into over 7 million experiments across India just to estimate yields. Deep Learning (Image Analytics) can help to reduce crop cutting experiments and deliver speedy claim settlement. Freely available Field Collector Software Development Kit (SDK) is a perfect solution for getting real-time ground data for risk assessment supported by geo-tagged images, farm details such as acreage, sowing date, calculating damage area, etc.
The above technological enhancement would lead to:
- Speedier claim settlements
- End-to-end transparency in claims process
- Reduce fraudulent claims
- Gain farmers trust and confidence for more insurance participation
- Proactive insurers and government agencies to deal with contingencies
- Automate processes saving time, money and labour.
(The writer is principal industry consultant – Insurance, SAS India)