openalea.phenomenal.tracking.trackedPlant.TrackedPlant

class openalea.phenomenal.tracking.trackedPlant.TrackedPlant(snapshots)

Bases: object

Main class for leaf tracking

__init__(snapshots)

Parameters:

snapshots (list(TrackedSnapshot)) –

Methods

__init__(snapshots)	param snapshots:
get_ref_skeleton([nmax])	Compute a median skeleton {rank : leaf}.
growing_leaf_tracking()	Tracking of growing leaves over time.
load(segmentation_time_series)	param segmentation_time_series:
mature_leaf_tracking([gap, ...])	alignment and rank attributions in a time-series of sequences of leaves.
output()

get_ref_skeleton(nmax=15)

Compute a median skeleton {rank : leaf}. For each rank, the leaf whose vector is less distant to all other leaves from the same ranks is selected.

Parameters:

nmax (int) – max number of leaves considered at a given rank (to avoid old leaves which can have senescence)

growing_leaf_tracking()

Tracking of growing leaves over time. To use AFTER self.align_mature()

static load(segmentation_time_series)

Parameters:

segmentation_time_series (list) –

list of dict {‘time’: float,

’polylines_sequence’: list of polylines, ‘features_sequence’: list of {‘mature’: bool, ‘azimuth’: float,

’height’: float, ‘length’: float}

}

Return type:

TrackedPlant

mature_leaf_tracking(gap=12.0, gap_extremity_factor=0.2, start=0, w_h=0.03, w_l=0.004, align_range=None, rank_attribution=True)

alignment and rank attributions in a time-series of sequences of leaves. Step 1 : use a multiple sequence alignment algorithm to align the sequences. Step 2 (post-processing) : Detect and remove abnormal group of leaves ; final rank attribution.

Parameters:

• gap (float) – weight for pairwise sequence alignment

• gap_extremity_factor (float) – parameter allowing to change the value of the gap penalty for terminal gaps (terminal gap penalty = gap * gap_extremity_factor)

• start (int) – sequences are progressively added to the global alignment from sequences[start] to sequences[0], then from sequences[start + 1] to sequences[-1]

• align_range (int) – When adding a new sequence to the global alignment, only the already aligned sequences with a distance inferior or equal to this parameter in the sequences order are used for the alignment.

• w_h (float) – weight associated to insertion height feature in a leaf feature vector

• w_l (float) – weight associated to length feature in a leaf feature vector

• rank_attribution (bool) – choose if step 2 is done (True) or not (False)