This project from the Amritavishwa Vidyapeetham, Amritapuri IEEE Student Branch involves the design and development of an AI-based coconut dehusking machine.
The proposed machine is an AC-powered dehusking machine. The machine has a dehusking unit, a chassis and a power transmission mechanism. The dehusking unit removes the coconut husk uniformly from the coconut. It has a special roller mechanism with dehusking blades welded on it. The roller mechanism will be designed to easily penetrate the husk. The chassis will be designed in sheet metal to mount the rollers and the transmission mechanism. The transmission mechanism will have a motor, gear mechanism, and chains for the transmission of power to the rollers.
The electronics control system involves a master controller, a microcontroller, and a speed regulation unit. The master controller will be an Intel i7 PC for running the classification algorithm. A camera with a stand will be attached to the chassis to capture images while feeding the coconut into the machine. An AI-based model running in the mini-pc will detect the class of the coconut and it will decide the speed of operation. A new dataset will be created by collecting the images of coconuts of various maturity levels for training the AI model. Based on the speed given by the mini-pc, the microcontroller will give commands to the variable frequency drive to control the AC motor.
This project was made possible by $5,000.00 in funding from the IEEE Industry Applications Society (IAS), an EPICS in IEEE partner.