This package is a general random sample consensus (RANSAC) framework. For convenience, some data models (such as a straight line) are already provided. However, you are free to define your own data models to remove outliers from arbitrary data sets using arbitrary data models.
There are two main components to this package: the RANSAC estimator and a
data model. When calling the estimation function find_inliers
, you need to
specify the model to which you expect your data to fit.
A data model is class containing the model parameters and an error function
against which you can test your data. Each data model must implement the
interface defined by the Model
base class. In other words, you need to
implement the make_model
and calc_error
functions.
Additionally, you need to provide parameters for the RANSAC algorithm. These
parameters are contained in the RansacParams
class.