Cross-Scale Dataset

The cross-scale dataset is built from the Scott Reef 25 dataset (2009) provided by Australian Centre for Field Robotics (ACFR). The dataset contains 1000 data tuples to split for training and validation, and 200 data tuples for testing.