Welcome to Flamingo’s documentation!

The Flamingo toolbox is an open-source toolbox for image segmentation, classification and rectification. It is developed by the Department of Hydraulic Engineering of Delft University of Technology for coastal image analysis. The toolbox is built around the scikit-image, scikit-learn, OpenCV and pystruct toolboxes.

Flamingo is developed and maintained by:

Bas Hoonhout <b.m.hoonhout@tudelft.nl>
Max Radermacher <m.radermacher@tudelft.nl>

The toolbox can be found at http://github.com/openearth/flamingo.

Command-line tools

Several command-line functions are supplied with the toolbox for batch processing of large datasets. Each command-line function serves a specific part of the image analysis. See for more information the –help option of each command.

rectify-images

classify-images

calibrate-camera

File system

The toolbox uses a file system structure for the analysis of datasets. The filesys module takes care of any reading and writing of files in this file structure. Each dataset is stored in a single directory and can consist out of the following file types:

Image files
Any image file recognized by the system
Cropped image files
Names start with cropped_. A non-cropped version of the image file should exist.
Export files
Pickle files with data concerning an image. Each export file name has the following format: <image_name>.<key>.pkl. A special type of export file is the feature file. Not all features are written to a single export file, but they are subdivided into multiple export files depending on the feature block they belong to. The block name is added to the export file, just before the file extension.
Log files
Pickle files with data concerning the entire dataset. Log file names can have any name.
Model files
Pickle files with a trained model. Each model file is accompanied by a meta file. Each model file name has the following format: model_<model_type>_<dataset>_I<nr_of_images>_B<nr_of_blocks>_<timestamp>.pkl. The corresponding meta file has meta added to the name, just before the file extension.

Configuration

Only the very basic options of the toolbox are exposed through the command-line functions. For the full extent of options a configuration file is used. This configuration file is parsed by the config module. The module also supplies wrappers for the automated updating of a function call based on the configuration file used.

config.CLASSIFICATION_DEFAULTS = {'channels': {'enabled': True, 'methods': ['gabor', 'gaussian', 'sobel'], 'methods_params': {'frequencies': [0.05, 0.15, 0.25], 'sigmas': [1, 8, 15], 'thetas': [0.0, 0.785, 1.571, 2.356]}}, 'segmentation': {'remove_disjoint': True, 'extract_contours': False, 'enabled': True, 'method': 'slic', 'method_params': {}}, 'relative_location': {'sigma': 2, 'enabled': False, 'n': 100}, 'features': {'feature_blocks': 'all', 'enabled': True, 'blocks_params': {}}, 'score': {}, 'partition': {'n_partitions': 5, 'force_split': False, 'enabled': True, 'frac_test': 0.25, 'frac_validation': 0.0}, 'training': {'partitions': 'all'}, 'regularization': {'C': [0.1, 1.0, 10.0, 100.0, 1000.0, 10000.0], 'partition': 0}, 'general': {'model_type': 'LR', 'colorspace': 'rgb', 'class_aggregation': '', 'model_dataset': ''}}

Configuration constants for classification toolbox

config.get_function_args(fcn, cfg, sections=[])[source]

Get relevant function arguments given a configuration file

config.parse_config(sections=[])[source]

Wrapper for parsing config file for specific function call

config.read_config(cfgfile, defaults={'channels': {'enabled': True, 'methods': ['gabor', 'gaussian', 'sobel'], 'methods_params': {'frequencies': [0.05, 0.15, 0.25], 'sigmas': [1, 8, 15], 'thetas': [0.0, 0.785, 1.571, 2.356]}}, 'segmentation': {'remove_disjoint': True, 'extract_contours': False, 'enabled': True, 'method': 'slic', 'method_params': {}}, 'relative_location': {'sigma': 2, 'enabled': False, 'n': 100}, 'features': {'feature_blocks': 'all', 'enabled': True, 'blocks_params': {}}, 'score': {}, 'partition': {'n_partitions': 5, 'force_split': False, 'enabled': True, 'frac_test': 0.25, 'frac_validation': 0.0}, 'training': {'partitions': 'all'}, 'regularization': {'C': [0.1, 1.0, 10.0, 100.0, 1000.0, 10000.0], 'partition': 0}, 'general': {'model_type': 'LR', 'colorspace': 'rgb', 'class_aggregation': '', 'model_dataset': ''}})[source]

Read configuration file and update default settings

config.write_config(cfgfile, defaults={'channels': {'enabled': True, 'methods': ['gabor', 'gaussian', 'sobel'], 'methods_params': {'frequencies': [0.05, 0.15, 0.25], 'sigmas': [1, 8, 15], 'thetas': [0.0, 0.785, 1.571, 2.356]}}, 'segmentation': {'remove_disjoint': True, 'extract_contours': False, 'enabled': True, 'method': 'slic', 'method_params': {}}, 'relative_location': {'sigma': 2, 'enabled': False, 'n': 100}, 'features': {'feature_blocks': 'all', 'enabled': True, 'blocks_params': {}}, 'score': {}, 'partition': {'n_partitions': 5, 'force_split': False, 'enabled': True, 'frac_test': 0.25, 'frac_validation': 0.0}, 'training': {'partitions': 'all'}, 'regularization': {'C': [0.1, 1.0, 10.0, 100.0, 1000.0, 10000.0], 'partition': 0}, 'general': {'model_type': 'LR', 'colorspace': 'rgb', 'class_aggregation': '', 'model_dataset': ''}})[source]

Write configuration file

Example configuration

[channels]
enabled = True
methods = ["gabor", "gaussian", "sobel"]
methods_params = {"frequencies": [0.05, 0.15, 0.25], "sigmas": [1, 8, 15], "thetas": [0.0, 0.785, 1.571, 2.356]}

[segmentation]
remove_disjoint = True
extract_contours = False
enabled = True
method = slic
method_params = {}

[relative_location]
sigma = 2
enabled = False
n = 100

[features]
feature_blocks = all
enabled = True
blocks_params = {}

[score]

[train]
partitions = all

[partition]
n_partitions = 5
force_split = False
enabled = True
frac_test = 0.25
frac_validation = 0.0

[regularization]
c = [0.1, 1.0, 10.0, 100.0, 1000.0, 10000.0]
partition = 0

[general]
model_type = LR
colorspace = rgb
class_aggregation = 
model_dataset =

Acknowledgements

The Flamingo toolbox is developed at Delft University of Technology with support from the ERC-Advanced Grant 291206 – Nearshore Monitoring and Modeling (NEMO), STW Grant 12686 – Nature-driven Nourishments of Coastal Systems (NatureCoast), S1: Coastal safety and Deltares.

Indices and tables