PriBeL: A primary betel leaf dataset from field and controlled environment

Area of Science:

• Agricultural Science
• Computer Science
• Pharmacology

Background:

• Visual identification of plant health is crucial for agriculture and medicinal plant research, impacting economics and pharmacology.
• Betel leaf (Piper betle) cultivation is globally important, but inconsistent classification and quality assessment arise from environmental factors and handling variations.
• AI integration in precision agriculture and quality control requires well-structured, diverse datasets for effective system performance.

Purpose of the Study:

• To introduce a systematically curated dataset of betel leaves to address challenges in classification and quality assessment.
• To provide a standardized and accessible resource for machine learning applications in agriculture and herbal medicine.

Main Methods:

• A dataset of 1,800 high-resolution (1080x1080 pixels) images was created, capturing betel leaves in Healthy (Fresh), Diseased, and Dried conditions.
• Images were collected from Pune, India, under diverse natural and controlled conditions, including varied lighting, backgrounds, and orientations to ensure comprehensive representation.
• The dataset is designed to cover real-world variations in betel leaf appearance.

Main Results:

• The Betel Leaf Dataset comprises 1,800 images across three distinct conditions: Healthy, Diseased, and Dried.
• The dataset captures variations in lighting, background, and orientation, reflecting diverse real-world scenarios.
• The curated images are standardized and accessible for research purposes.

Conclusions:

• The Betel Leaf Dataset enhances agricultural research by supporting leaf classification and quality assessment studies.
• This resource facilitates better documentation and understanding of betel leaf characteristics.
• The dataset is valuable for machine learning applications in plant disease detection, precision agriculture, and automated quality control systems.