Abstract
Chili plants played an important role in human life, serving as a source of income for farmers, as a provider of employment, and as a source of vitamins and minerals for the community. Market demand for red chili continued to increase, encouraging seed producers to provide quality plant seeds. The requirements for selecting plant varieties were based on market demand (taste, color, appearance, size, etc.), high productivity, resistance to plant pest attacks, and suitability for planting in local ecosystem conditions. Based on this, a smart approach was needed to identify plant varieties to maintain seed purity. To facilitate and streamline leaf-based chili variety identification, a comprehensive dataset was compiled. This dataset, consisting of 3877 leaf images divided into 12 variety classes, aimed to determine which plants were parent seeds or seeds that had deviations from their varieties. Leaf images were collected from the BALITSA garden through observations of leaf growth from shoots to 20 days of plant age. Various strict steps were taken to ensure the quality of the dataset and increase its usefulness. Chili leaf images taken from various angles and having high resolution were designed to assist in the development of highly accurate models. By leveraging this curated dataset, it was possible to train a model for real-time leaf-based identification of chili varieties, which significantly helped in the timely identification of such conditions.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call