This Bengaluru agritech startup is revolutionising seed classification with vision-based intelligence

Founded in May 2020, Bengaluru-based agritech startup Agdhi Technologies Pvt. Ltd. is one such startup that leverages ML and computer vision for seed classification, thereby enabling the farming community to procure quality seeds for better yield.“Seed companies spend a good amount of money and effort in identifying genetic purity, parental purity, hybrid purity, varietal purity, etc., using various methods like ODV, germination, and grow out tests (GOT). These methods are human-centric and highly expensive,” Nikhil Das, Co-founder and CEO, Agdhi, says
Through its AI, computer vision, photometry, and radiometry technology, Agdhi offers efficient methods for seed classification and quality analysis to determine a seed’s hybrid purity, parental purity, germination percentage, etc.
The backstory :
Nikhil Das, Anshad Ameenza, Ashwin Kashikar, and Vaibhav Kashikar got the idea for Agdhi while on a trip to an agricultural field.“I realised that although the technology is advancing in agriculture, the penetration of electronics hardware-based intelligent solutions is not matured in areas like seed industry,” Nikhil shares.
Hailing from Kerala, Nikhil is a tech innovator and entrepreneur who has built and scaled various startups in India. He manages Agdhi’s overall operations and resources, hardware/embedded team, and finance activities.
With over 16 years of experience, Anshad has built successful teams on AI, product engineering, cloud, Big Data infrastructure, data centre, etc. He handles Agadhi’s technology roadmap, software division, product roadmaps, and techno-commercial marketing.
With experience in agriculture and exposure to seed breeding and seed development and production activities, Ashwin Manages the Agri R&D division, while Vaibhav manages the marketing activities. He earlier worked with the National Seed Association of India and Seed Industries Association of Maharashtra, among others.
The startup has a team size of 30 members, consisting of AI/ML, front-end, and back-end developers, UX/UI team, embedded developers, hardware design team, quality control team, purchase manager, and production support team.
Products :
Agdhi’s flagship product SeedVision — an AI-based seed phenotyping method — helps in classifying critical parameters of a seed in a few minutes.
With SeedVision, each seed is analysed for its phenotypic character with the help of ML algorithms that predicts the yield outcomes, analyses seed health insight, and take action based on the insights.
While standalone solutions take images of seed samples and provide insights regarding the quality of the seed, Agdhi’s computer vision and ML models screens seed quality automatically, where its colour, texture, size, and shape are extracted from images, and the unique signatures of the seeds are identified through various classifications.
This automated technique provides an efficient method for seed sorting compared to manual inspection, and the reports are accessible on a real-time basis.
Nikhil says SeedVision costs between Rs 40 and 45 lakh for a single unit based on the seed varieties selected.
“SeedVision is indirectly helping farmers in getting good seeds for cultivation and is also helping growers improve the number of cycles because of faster payment realisation from seed companies,” he says.
The startup is also working on another product Planto — a mobile platform, backed by the learnings from the crop level data mapped by a plant’s entire life cycle. — which it will roll out as a separate product in the coming years to farmers.
Planto will help farmers in the early detection of potential crop damage, and provide analysis and suggestions to farmers on the health of the crop.
The app will also connect them to the industry experts, thereby helping them diagnose the crops from one’s mobile.
Business model :
According to MarketsandMarkets, the Indian seed industry is over a $2 billion market and spends over 15 percent of revenue in R&D.
Nikhil says, typically, a company takes five to seven years in releasing a hybrid into the market. Using SeedVision during R&D, the tracking and prediction become more accurate, which helps seed companies and agricultural research organisations save two to three years once they plot the growth parameters into our system.
“We also plan to tie up with government and private agencies, which will help smaller players to avail our service at an affordable cost,” he adds.
Source – Yourstory