ShuoZheLi/SVRDD_YOLO

SVRDD Road Damage Detection

This repository is formatted for the Hugging Face Dataset Viewer. Each row contains an image and its road-damage object annotations.

Load the dataset

from datasets import load_dataset

ds = load_dataset("YOUR_USERNAME/YOUR_DATASET_NAME")
example = ds["train"][0]
print(example["image"])
print(example["objects"])

Local unzip usage

If you want the images as regular local files, unzip the split archives:

unzip train.zip -d train_images
unzip validation.zip -d validation_images
unzip test.zip -d test_images

Annotation schema

• —image: the road image rendered by the Dataset Viewer.
• —file_name: image path referenced by the metadata row.
• —region: original Beijing district folder.
• —image_id: original image filename without extension.
• —width, height: image size in pixels.
• —objects: all annotations for the image.
• —bbox: [x_min, y_min, width, height] in absolute pixel coordinates.
• —bbox_yolo: [x_center, y_center, width, height] normalized to [0, 1].
• —categories: integer class ids from the YOLO labels.
• —category_names: human-readable class names.

Classes

Dataset Summary

• —Images: 8000
• —Objects: 20804
• —Regions: Chaoyang, Dongcheng, Fengtai, Haidian, Xicheng

Split Counts

Class Distribution