Classification of wheat seeds using image processing and fuzzy clustered random forest

Abstract

A reliable and autonomic seed classification technique can overcome the issues of manual seed classification. It is a highly practical and economically vital need of the agriculture industry. The current techniques of machine learning and artificial intelligence allows the researchers to design a new data mining mechanism with higher accuracy. In this article, a new adaptive technique has been proposed using a digital image processing system (DIPS) and fuzzy clustered random forest (FCRF) techniques. The DIPS is used to extract the parameters such as area, perimeter, height, width, length of the groove and asymmetry coefficient. Further, FCRF model is applied to classify the wheat seeds based on these parameters in a time-efficient manner. The devised approach helps the agriculture industry for seed classification, separation of damaged seeds and controlling the quality of seeds based on grading policy. The experiment result demonstrates that the accuracy of the proposed technique is better than the existing wheat seed classification algorithm. The average performance gain of the proposed technique is up to 97.7%.