DeepLoc Global Positioning Dataset

DeepLoc is a large-scale urban outdoor positioning dataset. The dataset currently consists of a scene spanning an area of 110 x 130 meters, which the robot traverses multiple times in different driving modes.

In addition, the dataset provides pixel-wise semantic segmentation annotations of 10 categories for each image. The dataset is divided into two parts: training and testing, with the training set including 2,737 images and the testing set including 1,173 images. The dataset also contains global GPS/INS data and LiDAR measurements.

This dataset is very challenging for vision-based applications such as global localization, camera relocalization, semantic segmentation, visual odometry, and loop closure detection because it contains extensive lighting, weather variations, repetitive structures, reflections, and transparent glass buildings.