FOCUS

This repository contains the implementation of FOCUS, our noise-aware geospatial deep learning framework for predicting PFAS contamination.

Getting Started

Dependencies

1. conda create -n <environment-name> python==3.9

2. conda activate <environment-name>

3. Install torch (tested for >=1.7.1 and <=1.11.0) and torchvision (tested for >=0.8.2 and <=0.12). May vary with your system. Please check at: https://pytorch.org/get-started/previous-versions/.

   1. e.g.: pip install torch==1.11.0+cu115 torchvision==0.12.0+cu115 --extra-index-url https://download.pytorch.org/whl/cu115

4. pip install -U openmim

5. mim install mmcv-full==1.6.2 -f https://download.openmmlab.com/mmcv/dist/{cuda_version}/{torch_version}/index.html. Note that pre-built wheels (fast installs without needing to build) only exist for some versions of torch and CUDA. Check compatibilities here: https://mmcv.readthedocs.io/en/v1.6.2/get_started/installation.html

   1. e.g.: mim install mmcv-full==1.6.2 -f https://download.openmmlab.com/mmcv/dist/cu115/torch1.11.0/index.html

6. Follow the required packages as mentioned in requirements.txt

Repository Structure

File	Description
pretraining.py	Script to pretrain the model on the multichannel geospatial features. Includes setup for masked image modeling.
main.py	Script to finetune the pretrained model using task-specific labels. Supports training routine.
patch_extraction.py	Utility to extract aligned image and mask patches for model training. Handles multichannel rasters and label masks.
noise_masks.py	Script to create noise-aware label masks based on distance, flow, land cover, and sampling priors.

Dependencies

pillow, rasterio, scikit-learn, scipy, joblib, tqdm, openpyxl, geopandas, tensorflow, pandas, numpy, pyproj, toolz, dask, shapely, nomkl, matplotlib, fiona, keras,

Data

The original PFAS monitoring data are provided by [NRSA, EPA Great Lakes studies, and MPART].

To run the code, you will need:

• Multi-band geospatial rasters with:
  • Land cover (e.g., NLCD)
  • Distance-to-facility bands
  • Flow direction / hydrological features
• Point-level PFAS measurements converted to raster labels

Data Availability. The dataset used in this study is large in size and is therefore hosted externally. The data are available on Google Drive.

Status

This repository provides the core implementation used in the FOCUS framework. We plan to release expanded documentation and usage examples in a future update.