Research Projects

According to the World Bank's most recent report, nearly all low and middle-income countries are experiencing food price inflation for staples like rice and maize. Food and Agriculture organization (FAO) estimates that given the low-rate at which crops are being produced, we could experience a severe food crisis within the next ten years. So, our aim is to find scientifically effective methods using which our farmers can increase food production irrespective of climate change thus maximizing their income. Agriculture contributes significantly to the Indian economy, which is why the Government of India has launched numerous schemes for farmers like Pradhan Mantri Krishi Sinchayee Yojana (PMKSY) and Beej Gram Yojna (BGY), to provide farmers with high-quality seeds to boost output. However, these high-quality seed and grains never reach the end-user Indian farmers, because government provided good quality seeds are sold by unethical vendors in secondary market at higher price and the poor-quality seeds are given to farmers under the scheme. So, it is very essential to have a mechanism for estimating the seed, grain, crop quality and health for their high agricultural yield potential. So, the development of low-cost non-invasive crop and seed screening method is very essential and impactful for Indian agriculture. Prior experience in various imaging techniques, industrial product design experience and India’s agricultural need has motivated PI and his expert team to investigate and design the computational Spectro-temporal method and apparatus for exploring the relationship between Indian crop grains, leaves and their yield potential for smart agriculture. Multi-sensory compact apparatus development also includes experimentation and estimation of appropriate spectra in various bands to acquire data from upper epidermis and seed kernels. To achieve robust results, the project will focus on developing deep learning-based image enhancer, ROI detector and probability-based classifier for classifying the food grain and crop as normal, mid-yield or high-yield potential product. The computational algorithm will be optimized to be able to execute on normal PC, so the entire screening system can be installed easily in any government department. The project will also focus on working with agricultural university and horticulture expert for obtaining insights, real-world agricultural data for the algorithm validation and improvements. This project will pave a way for using AI in agriculture tailored to Indian local crops, grains and it will also open up possibilities for extending the method for early crop disease prediction. Novelty of the proposal lies in the design of the multi-sensory hardware system, estimation of the illuminant’s frequency response for Indian crops & seeds with peculiar pigmentation as well as in the development of AI based computational approach for the early stage crop’s yield potential estimation.