IOT and Deep Learning Assisted Smart Agriculture System for Plant Disease Identification

Abstract

Agriculture employed more than 50% of labours and contributed nearly 18% to country's GDP. India is the leading producers of Many crops, such as wheat, rice, lentils, cotton, peanuts and perishable products but in recent years, the percentage of people getting into agriculture has been reduced. Crop disease and incorrect maintenance of humidity and moisture is the major reason that makes new farmers frustrated. Developing a deep learning model, that can identify the crop disease with high accuracy, and using this model one can easily identify the crop disease which makes easier for them treat the crop with pesticide, so that we cant wait for the experts. Using the IOT architecture to find the humidity and soil moisture of the field so that the farmer can maintain the correct level of humidity and soil moisture through irrigation, sometimes extreme high and very low moisture and humidity leads to diseases. An enhanced classification model was suggested in this paper to classify leaf diseases. Dataset with more than 10000 images is used for image classification. Classification accuracy is achieved using AlexNet model and it came out to be the highest in comparison to other models, and IOT part using the Raspberry pi 3 kit with soil moisture and humidity and soil moisture sensor.