Identification of Wheat Grain Using Geometrical Feature and Machine Learning

Plant breeding is the science of modifying the attributes of plants to achieve desired traits. It has improved the quality of nutritional material. Plant breeding aims to develop crop varieties that have distinctive favourable attributes for various agricultural applications. Wheat is an essential cereal crop for both the human and animal. The variety of wheat grain identification is a challenging task for people who are working in the field of plant breeding. This paper aims to identify the three wheat grain varieties, such as Kama, Rosa, and Canadian. Here, the six geometrical features such as area, perimeter, length, width, compactness, and asymmetry coefficient are considered from the wheat grain X-ray images. Again, three classifiers, namely Support Vector Machine (SVM), K-nearest neighbours (KNN), and naïve Bayes with their paradigms, are evaluated. The highest classification accuracy is achieved by fine-KNN. i.e., 94.3%. The AUC value is 0.93.