ML Based Quality Assurance System for Use in Areca Sheath Products Manufacturing

Ensuring on quality mainteanece, sustainability and development in Areca nut Sheath Product Manufacturing Areca sheath based dinnerware manufacturers are typically tiny or mini industries, employing mostly unskilled and a few semi-skilled labours. The manufacturing process consists of curing areca sheaths in water overnight, passing them through hot pneumatically operated presses with moulds to get different sized dinnerware products. Some of the challenges in manufacturing include unprepared sheaths and presence of fungus in parts of sheaths. The integrated AI based solution addresses challenges.

“Tiny and Micro industries manufacturing areca sheath products are sensitive to capital investments. Low-cost technology interventions that provide cost effective solutions in addressing quality issues and wastage control are likely to boost their bottom line in business. The integrated test rig with ML based solution that has been developed would likely improve their competitiveness in domestic and overseas markets”

Dr. Sridharakumar Narasimhan

Professor Dept. of Chemical Engineering, IIT-Madras

Key Features of the Project

Features of the integrated solution: A low-cost custom rig has been fabricated, with a flat bed, lighting, clamps, Raspberry PI microprocessor, camera and display monitor. The system indicates the process flow via a display monitor in vernacular language. The image of the sheath is captured. Three probes measure the conductivity of the sheath in different locations of the sheath The moisture content results are displayed visually in the monitor screen. Then, the presence of fungus areas are displayed on the screen

Key Technology used in the solution: For the moisture measurement, multiple measurements are automatically made, using three probes, to allow for settling time of water in the sheath Fungus detection process involves ML based solution. Grayscale filtering. Pewitt edge detection to remove veins (that appear as fungus) DBSCAN clustering. Drawing convex hulls to identify fungus areas

Users of the solution include: Micro and tiny industries that manufacture areca sheath products Areca sheath product machine manufacturers