Machine Vision Based Detection of Foreign Material in Wheat Kernels Using Shape and Size Descriptors

Abstract

Emergence of Internet of Things (IoT) in recent years has facilitated on-line trading of agricultural produce in real-time environment. However, in order to have an effective on-line marketing, automated qualityinspection systemsare required invariably for estimating the market value of these products. Thus, increasing use of on-line marketing in recent years has necessitated the urgent need to develop such systems. Machine vision provides a suitable technology for the development of these quality inspection systems. The work reported in the present paper is a part of an integrated machine vision system developed for automated quality inspection of wheat and other cereal grains. Proposed machine vision module has been developed specifically for detection of foreign material in wheat kernels using neural classifier. The proposed neural classifier has been executed using shape and size based regional descriptors of wheat kernels using digital image processing. Maximum average accuracy of more than 98.5% has been achieved in the present work. Results of present investigations are quite promising. The proposed machine vision module has potential future for IoT (Internet of Things) enabled on-line marketing of agriculture produce in real time environment.