Tobacco being one of the perennial crops that are the major source of forex in Malawi. But due to the decrease in agricultural extension workers specializing in Tobacco this paper proposes a remedy for tobacco disease detection and grading using a convolution neural network (CNN). This paper proposes the use of CNN as a machine learning algorithm for the training a system to detect tobacco diseases while grading the tobacco. This paper also compares CNN algorithm with other neural network algorithms in terms of sensitivity, specificity and other parameters such that the CNN proves to be efficient and effective hence offering higher performance of the algorithms. This research work will be designing and developing a learning model, which will classify, grade and detect the disease in the tobacco plant leaf by using the convolutional neural network based on the type of the tobacco Leaf. As these convolutional networks work with a large number of images for the training of the model, real-time images will be collected from the reliable source such as Agricultural Research and Extension Trust (ARET) and Tobacco Auction Flow Point respectively.

Keywords : Image Processing, Convolutional Neural Network (CNN), Leaf Detection, Agricultural Research and Extension Trust (ARET), Tobacco Classification, Tobacco Grading