urbanlc.model.baseline.TMBaseline#
- class urbanlc.model.baseline.TMBaseline(*args, **kwargs)#
TMBaseline class for LCC with classical ML using data from Landsat 4 - 7 (TM sensor)
Methods
build_model(model_params)Build the specific model based on the provided model name and parameters.
Construct transformation maps and functions to map between class labels, indices, and names.
infer(img_paths)Perform inference on specified Landsat image paths and returns the inferred land cover maps.
load_model(checkpoint_path)Load a pre-trained model from a checkpoint file.
retrieve_images(img_paths[, gt_paths, ...])Retrieve and preprocess Landsat images and ground truth land cover data from specified paths for training/inference.
save_model(data)Save the current model and related information to a checkpoint file.
train(img_paths, gt_paths[, enable_cv, ...])Train the model using specified Landsat image and ground truth land cover paths, with optional cross-validation.
transform_pipeline(img)Transform the input image using specified transformations.
update_transform_map(y_train)Update the transform map based on the provided training labels.
validate(img_paths, gt_paths)Validate the model using specified Landsat image and ground truth land cover paths.