Hyperspectral Zoo

Description

This a testing ground for implementing deep learning models for hyperspectral data.

Datasets

Datasets downloaded from here.

Dataset	# Bands	Image Size	# Classes	Sensor (spectral range)
Indian Pines	200	145 x 145	16	AVIRIS ($0.4 - 2.5 * 10^6$ nm)
Salinas	204	512 x 217	16	AVIRIS
Pavia Centre	102	1096 x 1096	10	ROSIS
Pavia University	103	610 x 610	10	ROSIS

Models

• 1D CNN - Hu et al. "Deep Convolutional Neural Networks for Hyperspectral Image Classification", Journal of Sensors, 2015
• HSI-CNN - Luo et al. "HSI-CNN: A Novel Convolution Neural Network for Hyperspectral Image", ICALIP 2018

Set-Up

Conda

# clone project
git clone https://github.com/abigailstone/Hyperspectral-Zoo
cd Hyperspectral-Zoo

# create conda environment and install dependencies
conda env create -f environment.yaml -n myenv

# activate conda environment
conda activate myenv

Log in to Weights & Biases from the command line to set up W&B logging.

Training

Train model with default configuration

# train on CPU
python src/train.py

# train using Salinas data 
python src/train.py data=salinas

You can override any parameter from command line like this

python src/train.py trainer.max_epochs=20 data.batch_size=64

See Hydra docs for more info.