Automated Plant Disease Detection and Diagnosis using Deep Learning, A Review

Authors

  • Arshad Ahmad Yatoo, Dr. Amit Sharma

Abstract

In order to ensure food security amid a growing global population and shrinking of land under cultivation, it becomes highly imperative to detect and identify crop diseases at the right time as they tend to be the major factors to reduce the quality and quantity of production. Moreover, agriculture is an important factor of a country’s economic strength as it offers a primary source of income to a significant percentage of its population. Every year farmers suffer huge losses of quality and quantity of yield due to different types of diseases. Since most of these diseases develop symptoms on plant leaves, different innovative techniques have been developed to detect and identify them by analysis of leaf images. These methods of disease detection and identification have proved highly effective and accurate in comparison to the optical observation of leaves through the naked eye. In this paper, we perform a study of some prominent research works to see how computers and electronics have entered into the agriculture/horticulture sector and are contributing to disease detection and diagnosis by leveraging the capabilities of computer vision and deep learning. This paper mainly focusses on disease detection through analysis of colour, shape, and texture of the diseased leaf symptoms using deep learning especially convolutional neural networks. A comparison of different studies has been made vis-a-vis their methodology, dataset, and accuracy. Moreover, this study also proposes future works that need to be undertaken in this field.

Published

2020-02-29

Issue

Section

Articles