Comparison of Different CNN architectures for Grape Leaf Disease Classification

Agriculturists are forced to develop new techniques that enhance production in order to feed the expanding population since the human population is continually growing while the amount of land accessible for food cultivation is remaining the same. With regard to this, maintaining the wellbeing of the crops in the field is essential. The work in this article explores the detection of diseased leaves from the healthy ones followed by identification of type of disease. Black rot, measles and blight infections present in grapevine plants are detected using Convolutional Neural Network (CNN). In this work, a novel Gabor filter based MobileNet architecture is developed for the classification of diseased leaves and the performance of the proposed and existing architectures is compared and analyzed.