Implementation of Hang et al. 2020 Hyperspectral Image Classification with Attention Aided CNNs for tree species prediction.

├── conf                   # Config files for model training and evaluation
├── data                   #  Location to place data for model reading. Most data is too large to be in version control, see below
├── DeepTreeAttention                   # Source files
├── experiments                    # Model training and SLURM multi-gpu cluster experiments with comet dashboards 
├── models                    # Trained snapshots
├── docs                   #
├── tests                    # Automated pytest tests
├── www                   # repo images
├── LICENSE
└── README.md
└── environment.yml # Conda Environment for model training and tests

• Data pipeline to predict species class for a DeepForest bounding box (https://deepforest.readthedocs.io/) for NEON Woody Veg Data
• Data pipeline to predict species class for a bounding box with weakly learned labels from random forest
• Training Pipeline for Hyperspectral DeepTreeAttention Model
• Add site metadata
• Training Pipeline for RGB DeepTreeAttention Model (technically works but is ineffective)
• Learned fusion among data inputs
• Autoencoder for outlier detection

This repo is being tested as an open source project on comet_ml. Comet is a great machine learning dashboard. The project link is here. Major milestones will be listed below, but for fine-grained information on code, model structure and evaluation statistics, see individual comet pages. To recreate experiments, make sure to set your own comet_ml api key by creating a .comet.config in your home directory (see https://www.comet.ml/docs/python-sdk/advanced/ ).

• Hyperspectral only

See conf/tree_config.yml for training parameters.

The field data are from NEON's woody vegetation structure dataset. A curated .shp is found at data/processed/field.shp which contains species labels and utm coordinates of each tree stem

To generate training data from existing shapefiles of deepforest predictions

python experiments/Trees/generate.py

To generate new deepforest boxes, you will need to create a seperate conda environment. DeepForest requires tensorflow <2.0 where this repo is >2.0. The requirements are otherwise the same. To generate boxes see

python experiments/Trees/prepare_field_data.py

After creating training data the main entry point is

python experiments/Trees/run.py

The original network can be cited:

• Hang, Renlong, Zhu Li, Qingshan Liu, Pedram Ghamisi, and Shuvra S. Bhattacharyya. 2020. “Hyperspectral Image Classification with Attention Aided CNNs,” May. http://arxiv.org/abs/2005.11977.

This repo can be cited on Zenodo once a release is created.