An automatic IoT based crop disease detection technique for early warning system using hybrid soft computing techniques

Abstract

The Indian economy is directly dependent on agricultural production. As the world grows and moves rapidly, identifying crop leaf diseases plays an important role in agriculture. Most of them are related to automation, and you also need to maintain the crop leaf with a little automation, and each crop has a specific need. To survive, it must be met, so companies need to develop systems that can communicate with their users. Currently, precision agriculture is being introduced to increase yields using the latest cutting-edge agriculture technology. An automatic disease detection system used to make instant and accurate decisions about plant diseases for farmers. This speeds up the diagnostic process. But in recent episodes, the show seemed a bit out of focus. One of these methods does not work well with the most important method, system diagnostics. In this paper, we propose an IoT based disease detection technique for crop using hybrid soft computing techniques (CDD-HSC). First, we segment the disease area from test leaf image using improved sunflower optimization (ISO) algorithm which is an important aspect for disease classification. Second, we introduce a multi-swarm snake optimization (MSSO) algorithm for optimal feature selection among multiple features in feature extraction stage. Then, we illustrates coach-learning induced capsule neural network (CL-CNN) for diseases classification in crop leaf with multi-classes. IoT concept used to transfer classification results to the corresponding former through mobile for immediate prevention in crop leaf diseases detecting, which limits the unwanted human delay. Finally, the performance of proposed CDD-HSC technique can analyze with different datasets and the results should sows the effectiveness of proposed method over existing methods in terms of accuracy, precision, F-measure and precision.