Do we really need deep CNN for plant diseases identification?

Timely identification of plant diseases plays crucial roles in the management and decision-making to protect the agricultural yield and quality. In this research field, there have been so many efforts focused on deep learning, namely deep CNN. The CNN is powerful and essential for image processing; however, do we really need deep CNN for plant diseases identification, cannot the shallow CNN extract enough information? We proposed two methods namely SCNN-KSVM (Shallow CNN with Kernel SVM) and SCNN-RF (Shallow CNN with Random Forest) to solve this confusion. The comparison experiments with other deep learning models were carried out on three different datasets. The results show that the SCNN-KSVM and SCNN-RF outperform other pretrained deep models on the indicators of precision, recall, and F1-score, with fewer parameters. The combination of shallow CNN and classic machine learning classification algorithm is a positive attempt to deal with the plant diseases identification in a simple manner.