Evaluating the Effectiveness of Machine Learning and Computer Vision Techniques for the Early Detection of Maize Plant Disease

Abstract

Monitoring plant growth is a crucial agricultural duty. In addition, the prevention of plant diseases is an essential component of the agricultural infrastructure. This technique must be automated to keep up with the rising food demand caused by increasing population expansion. This work evaluates this business, specifically the production of maize, which is a significant source of food worldwide. Ensure that Mazie's yields are not damaged is a crucial endeavour. Diseases affecting maize plants, such as Common Rust and Blight, are a significant production deterrent. To reduce waste and boost production and disease detection efficiencies, the automation of disease detection is a crucial strategy for the agricultural sector. The optimal solution is a self-diagnosing system that employs machine learning and computer vision to distinguish between damaged and healthy plants. The workflow for machine learning consists of data collection, data preprocessing, model selection, model training and testing, and evaluation.

Code and data availability

The paper uses a Kaggle maize leaf disease dataset (derived from PlantVillage/PlantDoc) and a ResNet50 transfer-learning model, but provides no authors' public URL, deposit, or availability statement for the dataset subset, code, or trained model. The Kaggle dataset is referenced generically without a link, and no code

No evidence-backed public reproduction asset is currently recorded.