SuperPoint SuperGlue TensorRT

Accelerate SuperPoint and SuperGlue with TensorRT.

Demo

• image pairs are from the freiburg_sequence.

Baseline(ToDo)

Image Size: 320 x 240	RTX3090	RTX3080	Quadro P620	Jetson Nano	Jetson TX2 NX
SuperPoint (247 points)
SuperPoint (237 points)
SuperGlue (256 dims)

• Compare the performance and speed with the offical SuperGlue.

Docker(Recommand)

docker pull yuefan2022/tensorrt-ubuntu20.04-cuda11.6:latest
docker run -it --env DISPLAY=$DISPLAY --volume /tmp/.X11-unix:/tmp/.X11-unix --privileged --runtime nvidia --gpus all --volume ${PWD}:/workspace --workdir /workspace --name tensorrt yuefan2022/tensorrt-ubuntu20.04-cuda11.6:latest /bin/bash

Environment Required

• CUDA
• TensorRT
• OpenCV
• EIGEN
• yaml-cpp

Convert Model(Optional)

python convert2onnx/convert_superpoint_to_onnx.py --weight_file superpoint_pth_file_path --output_dir superpoint_onnx_file_dir
python convert2onnx/convert_superglue_to_onnx.py --weight_file superglue_pth_file_path --output_dir superglue_onnx_file_dir

Build and Run

git clone https://github.com/yuefanhao/SuperPointSuperGlueAcceleration.git
cd SuperPointSuperGlueAcceleration
mkdir build
cd build
cmake ..
make
./superpointglueacceleration  ../config/config.yaml ../weights/ ${PWD}/../image/image0.png ${PWD}/../image/image1.png

If you modified the image size in the config file, you must delete the old .engine file in the weights dir.