Agrograde, a startup focused on autonomous technologies for fresh produce, has introduced an AI-powered optical sorting system designed specifically for the arecanut industry. The first commercial deployment of this technology was installed at Vikas Traders in Mangaluru, according to The Hindu BusinessLine. The system is intended to address persistent challenges in the arecanut supply chain, specifically the need to maintain consistent quality while processing large volumes of nuts.
The Hindu BusinessLine reported that the system uses computer vision and artificial intelligence to inspect individual arecanuts, identifying external defects that conventional colour-sorting machines might miss. The technology captures 10 to 12 images per nut using patent-pending vector technology, with proprietary AI models analysing these images within milliseconds to separate rejected material from accepted produce. The company stated that the models were trained on lakhs of crop images and run on NVIDIA-powered high-performance computing infrastructure.
In terms of capacity, the system can process approximately one tonne of arecanut per hour, enabling the Mangaluru facility to handle close to eight tonnes daily. Before the installation, Vikas Traders relied on manual inspection, where a group of about 15 workers could handle roughly 40 to 45 bags daily. The automated system now offers a more predictable processing capacity and generates quality summaries for every lot, including data on size distribution and defect intensity.
The Hindu BusinessLine noted that the system initially recorded approximately 75 per cent sorting accuracy when commissioned. As more commercial data became available from operations, the AI models were updated, raising accuracy to roughly 89 per cent, with additional gains anticipated as the system learns from new data. Rakeshkumar Barai, co-founder of Agrograde, said the development took nearly five months and three product iterations to create what the company calls India’s first autonomous optical sorter for arecanuts.
Raj Padaliya of Vikas Traders said the company was seeking technologies that could make their warehouse operations autonomous and fast. He noted that they collaborated closely with Agrograde to test multiple versions and that the sorting process is now nearly fully autonomous. Padaliya described Vikas Traders as the first private firm in the arecanut sector to adopt such advanced AI technology, expressing amazement at the hardware development by Indian startups.