Coupling Structural Descriptors with a Novel Semantic Graph Matching Approach for LiDAR Loop Detection

1. Prerequisites

We tested our code on ubuntu 20.04 and ubuntu 22.04.

• Eigen (3.3.7)
• PCL (1.12)
• Ceres-solver (2.1.0)

You need to install these libraries from official guidance.'

2. Data

KITTI

You can download the point cloud dataset from the KITTI official website. In our experiments, we use the labels from the SegNet4D. For the convenience, you can download from here.

Loop pairs: we use the distance-based criteria from the SSC. You also can download from our link.

3. Usage

3.1 Install

git clone https://github.com/SAGE-11/SAGE.git
mkdir build
cd build
cmake ..
make -j8

3.2 Loop Closure Detection

• KITTI dataset (distance-based)

Modify config/config_kitti_graph.yaml

eval_seq:
  cloud_path: "xx/kitti/sequences/02/velodyne/" # your LiDAR scans
  label_path: "xx/SegNet4D_predicitions/kitti/sequences/02/predictions/" # semantic predictions from our link
  pairs_file: "../loop_data/pairs/pairs_kitti/neg_100/02.txt" # loop pairs
  out_file: "../out/kitti/02.txt"  # output file for evaluating

Then, you can run the .bin file following this:

cd /SAGE/bin
./eval_lcd_seq

you can find the output file in the SAGE/out/. for evaluating, you can run:

cd /SAGE/scripts
python pr_curve.py

License

This project is free software made available under the MIT License. For details see the LICENSE file.