3_5

Global feature matching via DBoW2

• uses KITTI dataset
• detects ORB features from 5 images
• Creates vocabulary tree using DBoW2
• Matches features from 5 images using DBoW2

---

How to build & run

Requirement: OpenCV4, DBoW2

(DBoW3, FBoW can be used as well - but API will be slightly different)

Local build

mkdir build && cd build
cmake ..
make -j
./bow_matching

Docker build

Requires base build

docker build . -t slam:3_5
docker run -it --env DISPLAY=$DISPLAY -v /tmp/.X11-unix/:/tmp/.X11-unix:ro slam:3_5

# Inside docker container
cd fastcampus_slam_codes/3_5
./build/bow_matching