sits_tempcnn: Train temporal convolutional neural network models

In sits: Satellite Image Time Series Analysis for Earth Observation Data Cubes

Train temporal convolutional neural network models

Description

Use a TempCNN algorithm to classify data, which has two stages: a 1D CNN and a multi-layer perceptron. Users can define the depth of the 1D network, as well as the number of perceptron layers.

This function is based on the paper by Charlotte Pelletier referenced below. If you use this method, please cite the original tempCNN paper.

The torch version is based on the code made available by the BreizhCrops team: Marc Russwurm, Charlotte Pelletier, Marco Korner, Maximilian Zollner. The original python code is available at the website https://github.com/dl4sits/BreizhCrops. This code is licensed as GPL-3.

Usage

sits_tempcnn(
  samples = NULL,
  samples_validation = NULL,
  cnn_layers = c(256, 256, 256),
  cnn_kernels = c(5, 5, 5),
  cnn_dropout_rates = c(0.2, 0.2, 0.2),
  dense_layer_nodes = 256,
  dense_layer_dropout_rate = 0.5,
  epochs = 150,
  batch_size = 64,
  validation_split = 0.2,
  optimizer = torch::optim_adamw,
  opt_hparams = list(lr = 0.005, eps = 1e-08, weight_decay = 1e-06),
  lr_decay_epochs = 1,
  lr_decay_rate = 0.95,
  patience = 20,
  min_delta = 0.01,
  verbose = FALSE
)

Arguments

samples	Time series with the training samples.
samples_validation	Time series with the validation samples. if the samples_validation parameter is provided, the validation_split parameter is ignored.
cnn_layers	Number of 1D convolutional filters per layer
cnn_kernels	Size of the 1D convolutional kernels.
cnn_dropout_rates	Dropout rates for 1D convolutional filters.
dense_layer_nodes	Number of nodes in the dense layer.
dense_layer_dropout_rate	Dropout rate (0,1) for the dense layer.
epochs	Number of iterations to train the model.
batch_size	Number of samples per gradient update.
validation_split	Fraction of training data to be used for validation.
optimizer	Optimizer function to be used.
opt_hparams	Hyperparameters for optimizer: lr : Learning rate of the optimizer eps: Term added to the denominator to improve numerical stability. weight_decay: L2 regularization
lr_decay_epochs	Number of epochs to reduce learning rate.
lr_decay_rate	Decay factor for reducing learning rate.
patience	Number of epochs without improvements until training stops.
min_delta	Minimum improvement in loss function to reset the patience counter.
verbose	Verbosity mode (TRUE/FALSE). Default is FALSE.

Value

A fitted model to be used for classification.

Note

Please refer to the sits documentation available in <https://e-sensing.github.io/sitsbook/> for detailed examples.

Author(s)

Charlotte Pelletier, [e-mail]

Gilberto Camara, [e-mail]

Rolf Simoes, [e-mail]

Felipe Souza, [e-mail]

References

Charlotte Pelletier, Geoffrey Webb and François Petitjean, "Temporal Convolutional Neural Network for the Classification of Satellite Image Time Series", Remote Sensing, 11,523, 2019. DOI: 10.3390/rs11050523.

Examples

if (sits_run_examples()) {
    # create a TempCNN model
    torch_model <- sits_train(samples_modis_ndvi, sits_tempcnn())
    # plot the model
    plot(torch_model)
    # create a data cube from local files
    data_dir <- system.file("extdata/raster/mod13q1", package = "sits")
    cube <- sits_cube(
        source = "BDC",
        collection = "MOD13Q1-6",
        data_dir = data_dir
    )
    # classify a data cube
    probs_cube <- sits_classify(
        data = cube, ml_model = torch_model, output_dir = tempdir()
    )
    # plot the probability cube
    plot(probs_cube)
    # smooth the probability cube using Bayesian statistics
    bayes_cube <- sits_smooth(probs_cube, output_dir = tempdir())
    # plot the smoothed cube
    plot(bayes_cube)
    # label the probability cube
    label_cube <- sits_label_classification(
        bayes_cube,
        output_dir = tempdir()
    )
    # plot the labelled cube
    plot(label_cube)
}

sits documentation built on Nov. 2, 2023, 5:59 p.m.