Polarimetric Full-Wavefield Coherent Lidar

We present a polarimetric full-wavefield coherent lidar system that simultaneously recovers depth, velocity, and polarimetric properties of a scene from a single acquisition. By treating the reflected light as a full-wavefield signal and estimating the per-pixel Jones matrix via maximum likelihood, our system jointly reconstructs 3D geometry, Doppler velocity, and polarization state — enabling passive material characterization alongside conventional ranging.

The example dataset and calibration matrices can be downloaded from Dropbox.

pip install -r requirements.txt

Dependencies: Python ≥ 3.8, PyTorch (CUDA recommended), numpy, scipy, matplotlib, py_pol, tqdm, pyyaml.

public/
├── data/
│   ├── calibration/
│   └── fan_cat/
├── results/
│   ├── calibration/
│   └── fan_cat/
├── pfwl/
│   ├── models/               # reconstruction & calibration models
│   └── utils/                # config I/O, analysis, visualization
├── calibration.py            # Step 1: system calibration
├── pfwl_reconstruction.py    # Step 2: reconstruction
└── requirements.txt

Estimates the forward/backward fiber Jones matrices from polarizer rotation measurements stored in Calibration_matrices.npz.

conda run -n PFWL python calibration.py
# or
/path/to/envs/PFWL/bin/python calibration.py

Outputs → results/calibration/:

• optimization_results.npz — calibrated system matrices
• compare_ellipses.png — predicted vs. measured Jones ellipses

Runs per-pixel ML-NLL optimization to jointly estimate depth, velocity, and Jones matrix for each pixel.

conda run -n FWL python fwl_reconstruction.py
# or
/path/to/envs/FWL/bin/python fwl_reconstruction.py

Outputs → results/fan_cat/:

• NLL_fan_cat_epoch50.npz — depth, velocity, lambda2, Jones matrices
• depth.png — depth map
• velocity.png — Doppler velocity map
• mueller.png — 4×4 Mueller matrix image

@article{du2026polarimetric,
  title     = {Polarimetric Full-Wavefield Coherent Lidar},
  author    = {Du, Dongyu and Xie, Andrew and Mirdehghan, Parsa and Buscaino, Brandon
               and Baek, Seung-Hwan and Kutulakos, Kiriakos N. and Lindell, David B.},
  journal   = {Optica},
  volume    = {13},
  number    = {6},
  pages     = {1174--1183},
  year      = {2026},
  doi       = {10.1364/OPTICA.592823},
  publisher = {Optica Publishing Group}
}