When a plant is growing, it is vulnerable to many diseases. One of the trickiest issues in agriculture is the early detection of plant diseases. The diseases must be detected early otherwise, they may negatively impact the overall production, lowering the farmers' profits. Many researchers have presented various cutting-edge frameworks relying on Neural Networks to address this issue. However, the majorities of these processes either have limited prediction performance or use massive amounts of input variables. In this system, a hybrid approach of Long Short Term memory (LSTM) and Convolution Neural Network (CNN) is used to predict and classify different types of plant leaf diseases automatically. With just 9,825 training parameters, the suggested system achieves 98.65% testing accuracy and 99.23% training accuracy. When comparison to other methods described in the literature, the proposed hybrid model calls for fewer training parameters. As a result, the time required for training this model for detecting the plant diseases automatically and the time needed to diagnose the plant disease using the trained mode both can decrease significantly. It helps to farmer to diagnose various diseases easily at an initial stage, so that they can go for different treatments and some preventive measures.
Key words: Neural Networks, Long Short Term memory, Convolution Neural Network, LSTM, CNN.
|