def outlier_filter(diff, f=3, perc=None, max_dz=100):
print("Removing outliers")
print("Initial pixel count:")
print(diff.count())
print("Absolute dz filter: %0.2f" % max_dz)
#Absolute dz filter
diff = np.ma.masked_greater(diff, max_dz)
print(diff.count())
if perc is not None:
diff = filtlib.perc_fltr(diff, perc)
else:
#diff = filtlib.sigma_fltr(diff, f)
diff = filtlib.mad_fltr(diff, f)
print(diff.count())
return diff
def main():
parser = getparser()
args = parser.parse_args()
fn = args.fn
if not iolib.fn_check(fn):
sys.exit("Unable to locate input file: %s" % fn)
#Need some checks on these
param = args.param
print("Loading input raster into masked array")
ds = iolib.fn_getds(fn)
#Currently supports only single band operations
r = iolib.ds_getma(ds, 1)
#May need to cast input ma as float32 so np.nan filling works
#r = r.astype(np.float32)
#Want function that checks and returns float32 if necessary
#Should filter, then return original dtype
r_fltr = r
#Loop through all specified input filters
#for filt in args.filt:
filt = args.filt[0]
if len(param) == 1:
param = param[0]
param_str = ''
if filt == 'range':
#Range filter
param = [float(i) for i in param[1:]]
r_fltr = filtlib.range_fltr(r_fltr, param)
param_str = '_{0:0.2f}-{1:0.2f}'.format(*param)
elif filt == 'absrange':
#Range filter of absolute values
param = [float(i) for i in param[1:]]
r_fltr = filtlib.absrange_fltr(r_fltr, param)
param_str = '_{0:0.2f}-{1:0.2f}'.format(*param)
elif filt == 'perc':
#Percentile filter
param = [float(i) for i in param[1:]]
r_fltr = filtlib.perc_fltr(r, perc=param)
param_str = '_{0:0.2f}-{1:0.2f}'.format(*param)
elif filt == 'med':
#Median filter
param = int(param)
r_fltr = filtlib.rolling_fltr(r_fltr, f=np.nanmedian, size=param)
#r_fltr = filtlib.median_fltr(r_fltr, fsize=param, origmask=True)
#r_fltr = filtlib.median_fltr_skimage(r_fltr, radius=4, origmask=True)
param_str = '_%ipx' % param
elif filt == 'gauss':
#Gaussian filter (default)
param = int(param)
r_fltr = filtlib.gauss_fltr_astropy(r_fltr,
size=param,
origmask=False,
fill_interior=False)
param_str = '_%ipx' % param
elif filt == 'highpass':
#High pass filter
param = int(param)
r_fltr = filtlib.highpass(r_fltr, size=param)
param_str = '_%ipx' % param
elif filt == 'sigma':
#n*sigma filter, remove outliers
param = int(param)
r_fltr = filtlib.sigma_fltr(r_fltr, n=param)
param_str = '_n%i' % param
elif filt == 'mad':
#n*mad filter, remove outliers
#Maybe better to use a percentile filter
param = int(param)
r_fltr = filtlib.mad_fltr(r_fltr, n=param)
param_str = '_n%i' % param
elif filt == 'dz':
#Difference filter, need to specify ref_fn and range
#Could let the user compute their own dz, then just run a standard range or absrange filter
ref_fn = param[0]
ref_ds = warplib.memwarp_multi_fn([
ref_fn,
],
res=ds,
extent=ds,
t_srs=ds)[0]
ref = iolib.ds_getma(ref_ds)
param = [float(i) for i in param[1:]]
r_fltr = filtlib.dz_fltr_ma(r, ref, rangelim=param)
#param_str = '_{0:0.2f}-{1:0.2f}'.format(*param)
param_str = '_{0:0.0f}_{1:0.0f}'.format(*param)
else:
sys.exit("No filter type specified")
#Compute and print stats before/after
if args.stats:
print("Input stats:")
malib.print_stats(r)
print("Filtered stats:")
malib.print_stats(r_fltr)
#Write out
dst_fn = os.path.splitext(fn)[0] + '_%sfilt%s.tif' % (filt, param_str)
if args.outdir is not None:
outdir = args.outdir
if not os.path.exists(outdir):
os.makedirs(outdir)
dst_fn = os.path.join(outdir, os.path.split(dst_fn)[-1])
print("Writing out filtered raster: %s" % dst_fn)
iolib.writeGTiff(r_fltr, dst_fn, ds)