In EOdal 0.1.0 the clip method was introduced in the Band class and, hence, also the RasterCollection class.
The previous implementation, however, did only support clipping to a bounding box around the feature, i.e., in the case of a irregularly shaped feature, also pixels outside the feature (but within the bounding box) were returned. Obviously, this is not desired if one is only interested in obtaining the actual feature pixel values are statistics thereof.
Therefore, a new keyword was introduced (full_bounding_box_only) which set to False by default, meaning that pixel values outside the feature boundaries are masked (similar to the reading of raster data). If one is interested in analysing all pixel values within the bounding box, full_bounding_box_only must be set to True.
Further small changes include the introduction of a clip_scenes method as a wrapper around the Band.cip() method at the SceneCollection level and a small fix in the default SCL classes for Sentinel-2 in the Sentinel2.mask_clouds_and_shadows method (now everything but classes 4 and 5 are masked by default, i.e., vegetation and bare soil).
In EOdal 0.1.0 the clip method was introduced in the
Band
class and, hence, also theRasterCollection
class.The previous implementation, however, did only support clipping to a bounding box around the feature, i.e., in the case of a irregularly shaped feature, also pixels outside the feature (but within the bounding box) were returned. Obviously, this is not desired if one is only interested in obtaining the actual feature pixel values are statistics thereof.
Therefore, a new
keyword
was introduced (full_bounding_box_only
) which set toFalse
by default, meaning that pixel values outside the feature boundaries are masked (similar to the reading of raster data). If one is interested in analysing all pixel values within the bounding box,full_bounding_box_only
must be set toTrue
.Further small changes include the introduction of a
clip_scenes
method as a wrapper around theBand.cip()
method at theSceneCollection
level and a small fix in the default SCL classes for Sentinel-2 in theSentinel2.mask_clouds_and_shadows
method (now everything but classes 4 and 5 are masked by default, i.e., vegetation and bare soil).This pull request closes #22