Crop diseases pose a serious threat to food security. Fast classification of is crucial to overcome this threat but there is a serious shortage of this integral infrastructure to facilitate it. There is a growing demand worldwide for smartphones and the current advances in computer vision and computer assisted analysis has provided a feasible approach for smartphone assisted crop disease analysis. By utilising a publicly available data set of images containing both pathological and healthy leaves gathered in different settings, we train a deep convolutional neural network to classify crop disease pairs out of 14 crop species and 26 diseases present in those subspecies. Therefore the method of training deep learning models on these publicly available image data sets gives us a clear-cut path towards the direction of smartphone assisted crop disease prognosis on a huge global scale.