Abstract

Accurately diagnosing plant diseases poses a significant challenge for farmers throughout the growth and production stages. In the realm of plants, a disease is characterized by any disruption in the normal physiological function that manifests in identifiable symptoms. Symptoms, in this context, serve as evidence of the disease's existence and are typically observed on leaves or stems. The causal agents of diseases are pathogens, which instigate these physiological disruptions. Given that pests or diseases are frequently manifested on leaves or stems, precise identification of plants, leaves, stems, as well as the timely detection of pests or diseases, their occurrence percentages, and symptomatic expressions are pivotal for successful crop cultivation. The consequences of diseases are profound, resulting in significant crop losses and subsequent financial setbacks. To address this issue, we propose an enhanced deep-learning model designed for the accurate diagnosis of plant leaf diseases. The methodology involves employing an algorithm to cluster sample images as an initial step, followed by inputting them into the improved deep learning model for disease diagnosis. This innovative approach aims to enhance the efficiency and precision of disease identification, contributing to more effective agricultural practices and mitigating financial losses associated with crop damage.