To build and Implementation of a Deep Learning Model for IoT-based Smart Farming

The Plant diseases are the primary cause of decreased agricultural farming productivity. Most farmers struggle to recognize and control plant diseases. Early forecasting of these diseases will therefore help farmers avoid further losses. Deep learning techniques are widely used in tasks involving recognition and classification these days. Even though, many researchers use DL models for plant disease detection, they still have issues like the necessity for substantial data sets, low accuracy, the significant processing overhead, and the risk of overfitting. In order to solve such issues, a unique DL model is introduced. Here, an IoT-based smart farming method is developed using deep Bidirectional LSTM and artificial coyote optimization for analyzing and detecting plant disease using the IoT-collected data and soil information. Here, the input will be gathered from Plant Village and Soil Fertility Prediction datasets. IoT sensors will be used on farms during the data collecting phase to gather information about plant diseases, which will then be saved in the cloud. The control center will send a request message to obtain data from the cloud when the experts need the sensor data on the fields. Following the acquisition of data, pre-processing will be used to extract features from the pre-processed image. Following that, Deep BiLSTM gets the extracted features and refines them via Artificial Coyote optimization. In order to enhance the deep BILSTM model’s ability to accurately detect plant diseases, artificial coyote optimization was created by combining the qualities of artificial rabbit optimization and coyote optimization. The study is conducted using PYTHON, and performance is assessed in terms of accuracy, sensitivity, and specificity to demonstrate the efficacy of the method in identifying plant diseases.