THE EFFECTIVE APPLICATION OF CROP DISEASE RECOGNITION SYSTEMS IN SMART AGRICULTURE

Abstract

Artificial Intelligence (AI) has played a crucial role in smart agriculture. The
application of this technology has brought significant benefits to the agricultural
industry, ranging from increased production efficiency to minimizing negative
environmental impacts. Current research utilizes features such as color, shape,
and texture of images. These features are used in various combinations or as
inputs for Deep Convolutional Neural Networks (DCNNs) for image classification.
In this study, we propose a system architecture aimed at enhancing the contrast
of input images and extracting boundary features, utilizing color information of
images as inputs for DCNNs to recognize plant diseases effectively. Simulation
experiments on the Cropped-PlantDoc database demonstrate that the system
works efficiently and is highly applicable in edge devices.