Deep-CNN for Plant Disease Diagnosis Using Low Resolution Leaf Images

Abstract

Plant disease diagnosis is critical in agriculture since diseases frequently restrict plant production receptivity. Physical strategies to identify plant diseases are mostly timely, objecting, and lengthy. As a consequence, agricultural automation with automated identification of plant diseases is widely preferred. Most of the implemented models could only identify diseases of a particular plant using high-resolution images, which is quite expensive from a farmer's position. Because of the variation in leaf colors, aspect ratios, and congested backgrounds, detecting plant disease by low-quality images is difficult. This paper explores an efficient plant disease identification model that combines multiple plant diagnoses for low-resolution images using deep convolutional neural networks (DCNNs). This system acquires a multilabel classification to classify both the plant type and the specific disorder simultaneously.