DCNN and LSTM based Plant Leaf Disease Detection Model For Edge IoT Device

Edge Computing is an incredibly practical and affordable alternative IoT device for cloud computing that uses a revolutionary methodology. Not only does edge processing reduce the cost of systems administration frameworks, but it also improves edge-cloud latency, which is essential for all critical applications. Plant diseases have a catastrophic impact on food production security and can lead to a sharp decline in the quantity and quality of agricultural products. In severe cases, plant diseases may continue to be the cause of partial grain failure. As a result, an automated method for diagnosing and detecting plant diseases is vitally needed for all agricultural data. In this paper, we have proposed a plant disease detection method based on transfer learning and tested the model’s performance on IoT edge device. Long-Short Term Memory (LSTM) is used in the proposed system to classify plant leaf diseases based on data extracted by CNN. An internal memory within the LSTM network allows it to learn from critical experiences with long-term conditions. Therefore, the combination of CNN and LSTM layout features significantly enhances classification. It is suggested to use a Deep Convolutional Neural Network based on a pre-trained MobileNetV1 feature extractor and an LSTM classifier. The proposed model’s performance is contrasted with that of other pre-trained models, including NASNetMobile and MobileNetV2. The proposed model outperformed other models, achieving a validation accuracy of 0.9892.