Optimized Plant Disease Prediction using CNN and Fertilizer Recommendation Engine

Abstract

A large portion of the Indian economy depends on agricultural productivity, and the impact of plant diseases can be significant. The consequences of a plant being affected by a disease can result in a considerable decrease in output, and the experiencing decline in financial losses is a unique consequence of a decline in both the caliber and amount of agricultural goods. To avoid a reduction in agricultural productivity and quantity, it is essential to recognize plant diseases. While enormous acres of crops are being monitored, plant disease detection is receiving an increasing amount of attention. Thus, this is possible by making use of image processing methods for plant disease diagnosis. The major limitation of the existing system is, it only predicts a plant disease, and this work not only detects the plant disease but also recommends a suitable fertilizer. The dataset used in the proposed model contains 16870 images and the model is implemented using a deeper Convolutional Neural Network (CNN). The level of accuracy accomplished by the model is 96% for fruit leaves and 89% for vegetable leaves.