Skip to content

Dataset · stochastic

Street view panoramic images with geolocation labels

stochastic/random_streetview_images_pano_v0.0.2

Tens of thousands of stitched 360-degree Google Street View panoramas with labels for coordinate determination from photos.

What's inside

The dataset contains from 10 to 100 thousand panoramic images with about a 360-degree view, collected from randomstreetview.com, with text location labels. The dataset is designed to train algorithms to determine location solely from the visual content of the image.

What you can do with it

  • Training visual geolocation models (determining country and place from photos)
  • Classification of street scenes and regional architectural features
Tasks
Image classification
Size
10K–100K rows
Downloads this month
3.2K
Likes
31

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("stochastic/random_streetview_images_pano_v0.0.2")