Image-Based Dataset for Identification of Common Diseases in Indian Tea Plants

Current dataset comprises images of Indian tea leaves collected directly from tea farms (Talap Town, Tinsukia District, India), intended to support machine learning and deep learning applications in tea plant  leaf health monitoring and disease detection. The current dataset is organized into three distinct folders: (i) 01-Tea_Leaf_Dataset containing raw, clean images (197 healthy and 219 diseased), (ii) 02-Pre processed Images comprising seven classical preprocessing techniques, and (iii) 03-Adv Pre processed with six advanced augmentation techniques. The preprocessed images were designed to optimize performance for real-time visual models such as YOLO and other deep learning frameworks.In the second folder, each original image was expanded into seven variants using the following transformations: image resizing (416\u00d7416), histogram equalization using CLAHE, Gaussian blur, image normalization, edge detection (Canny), and augmentation via flipping and rotation. The third folder includes six additional transformations designed to further enhance model robustness: grayscale conversion, black-and-white thresholding (binary), HSV-based random color jitter, gamma correction, random zooming, and random angle rotations of \u00b130\u00b0. After preprocessing, the dataset expanded to 1,379 healthy and 1,533 diseased images in each of the enhanced folders. Current dataset serves as a valuable resource for researchers and practitioners aiming to develop and benchmark models in precision agriculture and plant pathology using computer vision.