def run_pixel(X, Y, dataset_config, yatsm_config, px=0, py=0):
""" Run a single pixel through YATSM
Args:
X (ndarray): 2D (nimage x nband) feature input from ordinal date
Y (ndarray): 2D (nband x nimage) image input
dataset_config (dict): dict of dataset configuration options
yatsm_config (dict): dict of YATSM algorithm options
px (int, optional): X (column) pixel reference
py (int, optional): Y (row) pixel reference
Returns:
model_result (ndarray): NumPy array of model results from YATSM
"""
# Extract design info
design_info = X.design_info
# Continue if valid observations are less than 50% of dataset
valid = cyprep.get_valid_mask(Y[:dataset_config['mask_band'] - 1, :],
dataset_config['min_values'],
dataset_config['max_values'])
if valid.sum() < Y.shape[1] / 2.0:
raise TSLengthException('Not enough valid observations')
# Otherwise continue with masked values
valid = (valid * np.in1d(Y[dataset_config['mask_band'] - 1, :],
dataset_config['mask_values'],
invert=True)).astype(np.bool)
Y = Y[:dataset_config['mask_band'] - 1, valid]
X = X[valid, :]
if yatsm_config['reverse']:
# TODO: do this earlier
X = np.flipud(X)
Y = np.fliplr(Y)
yatsm = YATSM(X,
Y,
consecutive=yatsm_config['consecutive'],
threshold=yatsm_config['threshold'],
min_obs=yatsm_config['min_obs'],
min_rmse=yatsm_config['min_rmse'],
test_indices=yatsm_config['test_indices'],
retrain_time=yatsm_config['retrain_time'],
screening=yatsm_config['screening'],
screening_crit=yatsm_config['screening_crit'],
green_band=dataset_config['green_band'] - 1,
swir1_band=dataset_config['swir1_band'] - 1,
remove_noise=yatsm_config['remove_noise'],
dynamic_rmse=yatsm_config['dynamic_rmse'],
slope_test=yatsm_config['slope_test'],
lassocv=yatsm_config['lassocv'],
design_info=design_info,
px=px,
py=py,
logger=logger)
yatsm.run()
if yatsm_config['commission_alpha']:
yatsm.record = yatsm.commission_test(yatsm_config['commission_alpha'])
if yatsm_config['robust']:
yatsm.record = yatsm.robust_record
if yatsm_config['calc_pheno']:
ltm = pheno.LongTermMeanPhenology(yatsm, yatsm_config['red_index'],
yatsm_config['nir_index'],
yatsm_config['blue_index'],
yatsm_config['scale'],
yatsm_config['evi_index'],
yatsm_config['evi_scale'])
yatsm.record = ltm.fit(year_interval=yatsm_config['year_interval'],
q_min=yatsm_config['q_min'],
q_max=yatsm_config['q_max'])
return yatsm.record
def run_pixel(X, Y, dataset_config, yatsm_config, px=0, py=0):
""" Run a single pixel through YATSM
Args:
X (ndarray): 2D (nimage x nband) feature input from ordinal date
Y (ndarray): 2D (nband x nimage) image input
dataset_config (dict): dict of dataset configuration options
yatsm_config (dict): dict of YATSM algorithm options
px (int, optional): X (column) pixel reference
py (int, optional): Y (row) pixel reference
Returns:
model_result (ndarray): NumPy array of model results from YATSM
"""
# Extract design info
design_info = X.design_info
# Continue if valid observations are less than 50% of dataset
valid = cyprep.get_valid_mask(
Y[:dataset_config['mask_band'] - 1, :],
dataset_config['min_values'],
dataset_config['max_values']
)
if valid.sum() < Y.shape[1] / 2.0:
raise TSLengthException('Not enough valid observations')
# Otherwise continue with masked values
valid = (valid * np.in1d(Y[dataset_config['mask_band'] - 1, :],
dataset_config['mask_values'],
invert=True)).astype(np.bool)
Y = Y[:dataset_config['mask_band'] - 1, valid]
X = X[valid, :]
if yatsm_config['reverse']:
# TODO: do this earlier
X = np.flipud(X)
Y = np.fliplr(Y)
yatsm = YATSM(X, Y,
consecutive=yatsm_config['consecutive'],
threshold=yatsm_config['threshold'],
min_obs=yatsm_config['min_obs'],
min_rmse=yatsm_config['min_rmse'],
test_indices=yatsm_config['test_indices'],
retrain_time=yatsm_config['retrain_time'],
screening=yatsm_config['screening'],
screening_crit=yatsm_config['screening_crit'],
green_band=dataset_config['green_band'] - 1,
swir1_band=dataset_config['swir1_band'] - 1,
remove_noise=yatsm_config['remove_noise'],
dynamic_rmse=yatsm_config['dynamic_rmse'],
slope_test=yatsm_config['slope_test'],
lassocv=yatsm_config['lassocv'],
design_info=design_info,
px=px,
py=py,
logger=logger)
yatsm.run()
if yatsm_config['commission_alpha']:
yatsm.record = yatsm.commission_test(yatsm_config['commission_alpha'])
if yatsm_config['robust']:
yatsm.record = yatsm.robust_record
if yatsm_config['calc_pheno']:
ltm = pheno.LongTermMeanPhenology(
yatsm,
yatsm_config['red_index'], yatsm_config['nir_index'],
yatsm_config['blue_index'], yatsm_config['scale'],
yatsm_config['evi_index'], yatsm_config['evi_scale'])
yatsm.record = ltm.fit(year_interval=yatsm_config['year_interval'],
q_min=yatsm_config['q_min'],
q_max=yatsm_config['q_max'])
return yatsm.record
# Run model
if reverse:
_X = np.flipud(X)
_Y = np.fliplr(Y)
else:
_X = X
_Y = Y
yatsm = YATSM(_X, _Y,
consecutive=consecutive,
threshold=threshold,
min_obs=min_obs,
min_rmse=min_rmse,
test_indices=test_indices,
retrain_time=retrain_time,
screening=screening,
screening_crit=screening_crit,
remove_noise=remove_noise,
dynamic_rmse=dynamic_rmse,
lassocv=lassocv,
logger=logger)
yatsm.run()
breakpoints = yatsm.record['break'][yatsm.record['break'] != 0]
print('Found {n} breakpoints'.format(n=breakpoints.size))
if breakpoints.size > 0:
for i, bp in enumerate(breakpoints):
print('Break {0}: {1}'.format(
i, dt.fromordinal(bp).strftime('%Y-%m-%d')))
