DPCN: Deep Phase Correlation for End-to-End Heterogeneous Sensor Measurements Matching

This is the official repository for DPCN, with paper "Deep Phase Correlation for End-to-End Heterogeneous Sensor Measurements Matching" accpepted and presented in CoRL2020.

Dependencies

There are a few dependencies required to run the code. They are listed below:

System Environments:

Python >= 3.5

CUDA 10.1

CuDnn

Pip dependencies:

Pytorch 1.5.0

Torchvision 0.6.0

Kornia 0.3.1

Graphviz

Opencv

Scipy

Matplotlib

Pandas

TensorboardX

You can install these dependencies by changing your current directory to this DPCN directory and running:

pip install -r requirements.txt -f https://download.pytorch.org/whl/torch_stable.html (if using pip),

or manually install these dependencies if it makes you feel more comfortable and assured.

We would suggest you to setup the whole environment in Conda or Virtualenv so that your original system will not be affected.

Also, we are now working on the upgrade process to be competible with the lastes version of packages (e.g. torch==1.7.1, cuda11, and kornia0.4.1), coming soon!

Pre-trained Models:

Pre-trained Models can be downloaded from this link: https://drive.google.com/file/d/1GZ8hz3cfaBP7F7KEdQK7M_WcwOM1GF6z/view?usp=sharing

How to Run The Code

Step 0

Before you start to train or validate on the simulation dataset, you might want to choose one specific dataset type of the following three: Homogeneous, Heterogeneous, and Dynamic Obstacles. Demonstrations are shown as follows:

By default, the code is running on Heterogeneous dataset and you can change this by modifying following lines in "./data/simulation.py".

To decide whether they are heterogeneous, modify line 78~80:

# for heterogeneous image, comment out if you want them to be homogeneous rot = cv2.GaussianBlur(rot, (9,9), 13) kernel = np.array([[-1,-1,-1],[-1,9,-1],[-1,-1,-1]]) rot = cv2.filter2D(rot, -1, kernel)
To decide whether there are dynamic obstacles, modify line 83~85

# for dynamic obstacles, comment out if you dont want any dynamic obstacles arr[0,] = add_plus(arr[0], *plus_location4) arr = np.reshape(arr, (1, height, width)).astype(np.float32)

Training

If you want to train the DPCN network from scratch on simulation dataset then simply run:

python trainDPCN.py --simulation.

By default, this will train the network on the heterogeneous dataset with the batch size of 2, and will run on GPU. There are several settings you can change by adding arguments below:

Arguments	What it will trigger	Default
--save_path	The path to save the checkpoint of every epoch	./checkpoints/
--simulation	The training will be applied on a randomly generated simulation dataset	False
--cpu	The Program will use cpu for the training	False
--load_pretrained	Choose whether to use a pretrained model to fine tune	Fasle
--load_path	The path to load a pretrained checkpoint	./checkpoints/checkpoint.pt
--load_optimizer	When using a pretrained model, options of loading it's optimizer	False
--pretrained_mode	Three options: 'all' for loading rotation and translation; 'rot' for loading only rotation; 'trans' for loading only translation	All
--use_dsnt	When enabled, the loss will be calculated via DSNT and MSELoss, or it will use a CELoss	False
--batch_size_train	The batch size of training	2
--batch_size_val	The batch size of training	2
--train_writer_path	Where to write the Log of training	./checkpoints/log/train/
--val_writer_path	Where to write the Log of validation	./checkpoints/log/val/

Validating

If you are only interested in validating on the randomly generated simulation dataset, then you can simply run following lines based on the specific dataset type you chose in Step 0.

For Homogeneous sets:

python validate.py --simulation --only_valid --load_path=./checkpoints/checkpoint_simulation_homo.pt

For Heterogeneous sets:

python validate.py --simulation --only_valid --load_path=./checkpoints/checkpoint_simulation_hetero.pt

For Dynamic Obstacle sets:

python validate.py --simulation --only_valid --load_path=./checkpoints/checkpoint_simulation_dynamic.pt

Again, Pre-trained Models can be downloaded from this link: https://drive.google.com/file/d/1GZ8hz3cfaBP7F7KEdQK7M_WcwOM1GF6z/view?usp=sharing

Similarly, there are several options that you can choose when running validation, shown as follows:

Arguments	What it will trigger	Default
--only_valid	You have to use this command if you run validation alone	False
--simulation	The training will be applied on a randomly generated simulation dataset	False
--cpu	The Program will use cpu for the training	False
--load_path	The path to load a pretrained checkpoint	./checkpoints/checkpoint.pt
--use_dsnt	When enabled, the loss will be calculated via DSNT and MSELoss, or it will use a CELoss	False
--batch_size_val	The batch size of training	2
--val_writer_path	Where to write the Log of validation	./checkpoints/log/val/

Detecting and Infering:

This repository also provided a single pair detecting script so that you can see the result of DPCN directly. A few demos are given in the "./demo" directory including images pairs from both simulation dataset and Aero-Ground Dataset. You could customize the script detect.py to test the chosen images pair with relative pre-trained model given, and run this code below:

python detect.py

The results should be something like this:

Citation

If our source code could help you in your project, please cite the following:

@InProceedings{chen2020deep,
  title = 	 {Deep Phase Correlation for End-to-End Heterogeneous Sensor Measurements Matching},
  author =       {Chen, Zexi and Xu, Xuecheng and Wang, Yue and Xiong, Rong},
  booktitle = 	 {Proceedings of the 2020 Conference on Robot Learning},
  pages = 	 {2359--2375},
  year = 	 {2021},
  volume = 	 {155},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {16--18 Nov},
  publisher =    {PMLR},
  pdf = 	 {https://proceedings.mlr.press/v155/chen21g/chen21g.pdf},
  url = 	 {https://proceedings.mlr.press/v155/chen21g.html},
}

ZJU-Robotics-Lab / DPCN