Identification of Canola Seeds using Nearest Neighbor and K-Nearest Neighbor Algorithms

Abstract

Agriculture plays an important role on Pakistan economy. Canola is the major crop of Pakistan. There are different varieties of canola crop. It fulfills the requirement of oil. It is the difficult task to identify best canola seeds for sowing due to different varieties of canola seeds. In this paper we are try to introduce different machine learning approaches for classification of different canola seeds which provide opportunity to people or farmer to identify different canola seeds. Canola seeds verities implementing by the computer vision image processing tool. We have the 4 different varieties which names as Gobhi Sarson (A), Barassica comp (B), Sathri (C) and Rocket Herbof (D) canola seeds and take the images of canola seeds from these different varieties. Each variety has 10 images and we have total 10*4 =40 images of canola seeds. we take the train and test data results of all kinds of canola seeds. then train and test data results will be compare for pattern classification and apply the nearest neighbor and k-nearest neighbor algorithms for final classification in computer image processing tool. We achieved in nearest neighbor 85% and 76% average and k-nearest neighbor 90% and 73% average as a final pattern classification results. These are the best percentage for classification and provide more accuracy. These are important for farmer and other people for identify the different canola seeds.

Keywords:  Features, Pattern classification, nearest neighbor, k-nearest neighbor