In a bid to reduce the delay of crop insurance claim settlements and increase the accuracy of compensation due to farmers, a pilot project will test the use of technology to determine yield estimates at the panchayat level this summer.Technologies such as satellite and remote sensing data, unmanned aerial vehicles and artificial intelligence will be used to assess yield estimates without the need of time-consuming and laborious crop-cutting experiments, according to the parameters of the project issued. “Crop yield information is essential for the Pradhan Mantri Fasal Bima Yojana to work, but the number of CCEs needed for accurate determination of yields have increased multi-fold. PMFBY guidelines call for four CCEs at the gram panchayat level…The number of CCEs for production estimates itself may go up six to seven times,” says Shibendu Shankar Ray, director of the Mahalanobis National Crop Forecast Centre, which is overseeing the project. The current methodology of yield estimation is also affected by the lack of current year information at the time of planning of CCEs, thus affecting the precision of estimates.Encouraging results The Agriculture Ministry has tried several methods to deal with this situation, including smart sampling methods to improve the way CCEs are selected, as well as the use of technology to reduce the number of CCEs needed through accurate extrapolation methods. The pilot project to optimise CCEs through technology was carried out in 11 States in kharif season 2018 and in the ongoing rabi season 2018-19. “The results have been encouraging, and there is a proposal to roll out some of these technologies for certain crops in certain areas in the next season if all stakeholders agree,” he said.The new study, to be carried out in the kharif 2019, aims to directly estimate yields without using CCEs. The pilot project will focus on paddy, soybean, cotton, bajra, maize, sorghum and groundnut. Technologies to be used include satellite data, unmanned aerial vehicles, advanced intelligent crop simulation models, artificial intelligence and the Internet of Things. Final results are expected by February 2020.