The Udila Tea Leaves Dataset: Enabling Deep Learning Classification for Premium Tea Production
Description
This dataset contains a curated collection of tea leaf images sourced from Udila Tea Garden in Chittagong, Bangladesh, specifically designed to classify leaves based on their suitability for high-quality tea production. The dataset is categorized into two classes: "Tea-Making Leaf" and "Not Tea-Making Leaf" with a total of 747 samples. The "Good for Tea-Making" class includes fresh, vibrant leaves identified as ideal for producing flavorful, premium tea. In contrast, the "Not Good for Tea-Making" class contains leaves that are either aged, unhealthy, or otherwise unsuitable for quality tea production. Data Collection: Images were captured under controlled natural lighting to accurately reflect the visual characteristics of each leaf. "Good" leaves are typically new, fresh, and exhibit a desirable green color and healthy structure. "Not Good" leaves may show signs of aging, poor health, or other qualities that would compromise the tea flavor and quality. Applications: This dataset can be used for machine learning models aimed at quality assessment in tea production, enabling automated sorting and decision-making for optimal tea leaf selection in agricultural and commercial tea production processes.
Files
Steps to reproduce
Data Collection: Visit the chosen tea garden, Udila Tea Garden, Chittagong, Bangladesh, to collect leaf samples. Capture images of tea leaves in controlled natural lighting conditions to ensure consistent image quality. Use a high-resolution camera, ensuring the leaf occupies the majority of the frame for clearer analysis of its features. Categorize the images into two classes: Good for Tea-Making: Leaves that meet specific criteria for high-quality tea production, such as vibrant color, healthy texture, and lack of physical defects. Not Good for Tea-Making: Leaves that exhibit characteristics associated with suboptimal tea production, such as discoloration, damage, or deformities.