A Comparative Study on Machine Learning Based Classifier Model for Wheat Seed Classification

Abstract

AbstractSeed classification is a process of categorizing different varieties of seeds into different classes on the basis of their morphological features. Seed identification is further complicated due to common object recognition constraints such as light, pose and orientation. Wheat has always been one of the globally common consumed foods in India. A large number of wheat varieties have been cultivated, exported and imported all around the world. Enormous studies have been done on identifying crop diseases and classifying the crop types. In the present work, wheat seed classification is performed to distinguish the three different Indian wheat varieties by their collected morphological features and applied machine learning models to develop wheat variety classification system. The seed features used here are length of kernel, compactness, asymmetry coefficient, width of kernel, length of kernel groove, area and perimeter. The present work carried out with different classifiers such as Decision Tree, Random Forest, Neural Net, Nearest Neighbors, Gaussian Process, AdaBoost, Naive Bayes, Support Vector Machine(SVM) Linear, SVM RBF(SVM with the Radial Basis Function) and SVM Sigmoid with 2 K-fold cross validation. Also obtained the results using 5 fold and 10-fold Cross Validation.KeywordsAgricultureSeed classificationSVMNeural netGaussianK-fold cross validationRandom forestDecision treeClassifiers