A Multi-Plant Disease Diagnosis Method Using Convolutional Neural Network

Diagnosing the plant disease is crucial from perspective of agriculture, as diseases often limit plants’ production capacity. However, manual approaches to recognize are temporal, challenging, and time-consuming. Therefore, computerized recognition highly desired in field agricultural automation. Due recent improvement computer vision, identifying using leaf images a particular has already been introduced. Nevertheless, most introduced models can only diagnose specific plant. Hence, this chapter, we investigate an optimal identification model combining diagnosis multiple plants. Despite relying on multi-class classification, inherits multi-label classification method identify type parallel. For experiment evaluation, have collected data various online sources that included six plants, including tomato, potato, rice, corn, grape, apple. In our investigation, implement numerous popular convolutional neural network (CNN) architectures. The experimental results validate Xception well DenseNet architectures perform better tasks. Besides, found CNN architecture, skip connections, spatial convolutions, shorter hidden layer connectivity influence classification.