Plant Leaf Disease Detection Using CNN and Ensemble Classification

Abstract

The chief origin of economic progress in India is agronomics. Suitable are crops are chosen by the farmers throughout the year in every season depending on weather circumstances, nature of soil and economic value of the crop. The agronomic industries are looking forward for upgraded methods of production of food. In order to boost up the crop revenue and minimize the investment, the researchers have been working on new and latest technologies. One of the primary concerns in this area is the plant leaf disease. These infections not only reduce the yield but also a regress in the economy. So, for a country as India in which backbone of the economy, it becomes significantly important to design an efficient way to detect leaf diseases. The primary goal of this paper is to identify the leaf disease using the concept of ensemble learning of traditional classifier withconvolutional neural networks. The CNN extracts the required features from the images that are input into the model and then the classification is performed with the help of an ensemble of CNN and Random Forest(RF). These features aid the model in identifying the most appropriate label for the disease. It has been observed that our system has an average accuracy of approximate 96%.