Authors: Umer Khan
Abstract: Agriculture is a primary source of livelihood for a large population in India and remains essential for human survival and national economic development. However, variations in climate and local environmental conditions increase the risk of crop diseases, which can significantly reduce yield and quality. In many cases, the earliest symptoms of plant infections appear on leaves and, if not detected in time, the disease can spread throughout the plant and across the field, leading to major production losses. Early and accurate identification of plant diseases is therefore critical for reducing agricultural losses and improving the quality of farm produce. Since manual inspection is difficult and time-consuming due to the large number of plants in cultivation areas, automated disease detection methods are increasingly required. This work proposes an AI-based software approach for detecting leaf infections to enable fast and reliable diagnosis, followed by evaluation and actionable insights to prevent large-scale crop damage. The proposed framework involves key steps including image dataset collection, image preprocessing, feature extraction/selection from leaf images, model evaluation, and disease classification. The overall goal is to support timely disease management and improve crop productivity and profitability.
