Wheat Powdery Mildew Image Dataset

The dataset consists of wheat leaf images collected using a mobile phone, specifically focused on capturing diseased areas. These images are annotated with labels indicating the presence of diseases, suitable for training and testing a YOLO (You Only Look Once) object detection model. The labeled dataset aims to enable the model to accurately identify and classify diseased regions in similar images.

Description of the data and file structure

1. Images Folder:

Structure: The Images folder contains three subfolders: train, test, and val. Each subfolder contains image files in formats such as JPEG or PNG.

train: Contains images used for training the YOLO model.

test: Contains images used for testing the model's performance.

val: Contains images used for validating the model during training, helping to tune hyperparameters and prevent overfitting.

2. Labels Folder:

Structure: The Labels folder mirrors the structure of the Images folder, with subfolders named train, test, and val. Each subfolder contains YOLO-format label files.

train: Contains label files corresponding to the training images.

test: Contains label files for the testing images.

val: Contains label files for the validation images.

3. YOLO Label Format:

Contents: Each label file corresponds to an image and contains information in the YOLO format, which includes:

Class ID: An integer representing the class label (e.g., a specific disease).

Bounding Box Coordinates: Four numbers representing the center x, center y, width, and height of the bounding box, all normalized between 0 and 1.

File Naming Convention: The label files are named identically to their corresponding images, except for the file extension (e.g., image1.jpg and image1.txt).

4. Usage and Application:

Model Training and Validation: The dataset can be used to train and validate YOLO object detection models. The clear separation into train, test, and validation sets supports robust model evaluation and helps prevent data leakage.

Model Testing: The test set is used to evaluate the model's performance on unseen data, providing an unbiased measure of its generalization capability.

Model Tuning: The validation set helps fine-tune model parameters and assess performance during the training process.