Module for calculating doubly logistic approximation.
- eolearn.features.doubly_logistic_approximation.doubly_logistic(middle, initial_value, scale, a1, a2, a3, a4, a5)
Function that is passed to scipy.optimize
- class eolearn.features.doubly_logistic_approximation.DoublyLogisticApproximationTask(feature, new_feature='DOUBLY_LOGISTIC_PARAM', initial_parameters=None, valid_mask=None)
EOTask class for calculation of doubly logistic approximation on each pixel for a feature. The task creates new feature with the function parameters for each pixel as vectors. :param feature: A feature on which the function will be approximated :type feature: (FeatureType, str) :param new_feature: Name of the new feature where parameters of the function are saved :type new_feature: (FeatureType, str) :param initial_parameters: Initial parameter guess :type initial_parameters: List of floats length 7 corresponding to each parameter :param valid_mask: A feature used as a mask for valid regions. If left as None the whole patch is used :type valid_mask: (FeatureType, str) or None
Stores initialization parameters and the order to the instance attribute init_args.
eopatch – Input eopatch with data on which the doubly logistic approximation is computed
Output patch with doubly logistic approximation parameters