The implementation of LocNet back-end is based on caffe, and this page contains the caffe models of LocNet network.
In the models folder, for example, kitti_delta_range means that this model is used in KITTI dataset with Velodyne HDL-64E, with the Δ-r representations.
The input of the network is the image-like representations of LiDAR data, which can be generated using LocNet_frontend.
If you use our code in an academic work, please cite the following paper:
@article{yin20193d,
title={3D LiDAR-Based Global Localization Using Siamese Neural Network},
author={Yin, Huan and Wang, Yue and Ding, Xiaqing and Tang, Li and Huang, Shoudong and Xiong, Rong},
journal={IEEE Transactions on Intelligent Transportation Systems},
year={2019},
publisher={IEEE}
}
or other related conferences:
title={LocNet: Global localization in 3D point clouds for mobile vehicles}
title={Efficient 3D LIDAR based loop closing using deep neural network}