
Arvind kumar
Former Vice- ChancellorRani Lakshmi Bai Central Agricultural Uni., Jhansi, U.P., INDIA

Submitted
2025/12/15First Online
2026/02/25doi
10.18805/ag.DF-833Cite article - Kunal D. Jadhav, Neha Pokhriyal, Taniya Sengupta Rathore, M.B. Abhishek, Niveditha Pallerla. (2026). Plant Disease Pathology: Causes, Machine Learning-based Detection and Sustainable Management Strategies. Agricultural Science Digest. 46(5): 891–898. doi: 10.18805/ag.DF-833
Background: Plant diseases are a major challenge for global food production. They lead to significant reductions in crop yield and quality. Fungi, bacteria, viruses and nematodes are common pathogens that attack plants. These diseases not only affect food security but also cause economic losses worldwide. Estimates show that 20-30% of global crop yields are lost annually due to plant diseases. Early detection is necessary to minimize losses and protect crops. Traditional detection methods rely on field observation and laboratory tests. These techniques are time-consuming, labor-intensive and may not be practical for large-scale monitoring. Modern tools, including imaging technologies and machine learning, are emerging as effective solutions. They offer rapid, accurate detection and classification of plant diseases.
Methods: This paper reviews plant disease detection techniques with a focus on classification systems and diagnostic tools. The classification is based on disease incidence, mode of spread, symptoms, host parts affected and causative agents. A literature search was performed using databases such as Scopus, Web of Science and Google Scholar. Keywords included “plant disease detection,” “machine learning,” “AI in agriculture,” and “disease management.” Studies from 2005 to 2024 were considered. Priority was given to research discussing image-based diagnosis, hyperspectral imaging and machine learning models. Relevant articles were analyzed for methods, performance and limitations.
Result: Machine learning-based models show strong potential for disease detection. Convolutional neural networks (CNNs) are widely used for image classification tasks. Hyperspectral imaging and sensor-based systems improve accuracy. However, limitations exist. Models struggle with dataset imbalance, varying environmental conditions and real-field application. More diverse datasets and field validation are needed. Explainable AI models are also lacking.
Distinguished scholars and subject experts committed to editorial excellence, scientific integrity and rigorous peer review.
View Full Board