Bounding Box Dataset Augmentation for Long-range Object Distance Estimation
Marten Franke, Vaishnavi Gopinath, Chaitra Reddy, Danijela Ristić–Durrant, Kai Michels · 2021
Autonomous long-range obstacle detection and distance estimation plays an important role in numerous applications such as railway applications when it comes to locomotive drivers support or developments towards driverless trains. To overcome the problem of small training datasets, this paper presents two data augmentation methods for training the ANN DisNet to perform reliable long-range distance estimation.