The HelmetViolations dataset focuses on identifying and classifying whether a motorcycle rider is wearing a helmet from a top-down view, as well as detecting motorcycle license plates. This dataset is exported via Roboflow and is intended for object detection tasks based on YOLOv9, making it particularly suitable for projects that improve road safety and enforce helmet laws through automated systems.

The dataset contains a total of 1,004 images, annotated in YOLOv9 format, and includes 3 categories: license plate (Plate), helmet (WithHelmet), and no helmet (WithoutHelmet). The training set has 363 images (original + enhanced); the validation set has 53 images; the test set is included in the export for model evaluation.

The images were resized to 640×640 resolution, auto-orientation was applied (EXIF removed), and they were converted to grayscale images (CRT fluorescence effect). To increase diversity and improve the generalization ability of the model, the following augmentation operations were applied to each source image: random rotation between -15° and +15°, and horizontal flipping with a probability of 50%.