Performance Analysis among Different Classifier Including Naive Bayes, Support Vector Machine and C4.5 for Automatic Weeds Classification
Weeds are often one of the biggest problems encountered by farmer in conventional agriculture. Maximum productivity of crops can be achieved by proper weeds management. Applying excessive herbicide uniformly throughout the field has an adverse effect on the environment. An automated weed control system which can differentiate the weeds and crops from the digital image could be a feasible solution for this problem. This paper demonstrates Naïve Bayes, SVM (Support Vector Machine) and C 4.5 classification algorithm for classifying the weeds and investigates the performance analysis among these three algorithms. In this study 400 sample images over five species were taken where each and every species contains 80 images. The result has shown that Naïve Bayes classification algorithm achieve the highest accuracy (99.3%) among these three classifier.
