def mark_useg(mdata, map, pt, mark, structure):
"""
Mark upward-connected region on segment map starting at node pt
Parameters:
* mdata Masked data array
* map Array mapping out segments
* pt Index of starting node
* mark Integer to mark map with.
Returns nothing but modified mdata mask and map
"""
if mdata.mask[pt] == True:
return
else:
map[pt] = mark
mdata[pt] = ma.masked
Q = [pt]
while Q:
pt = Q.pop(0)
v = mdata.data[pt]
# Check all neightbors
for new_pt in neighbors(pt, mdata.shape, structure):
if mdata.mask[new_pt] == False and mdata[new_pt] > v:
Q.append(new_pt)
map[new_pt] = mark
mdata[new_pt] = ma.masked
return
def mark_useg(mdata, map, pt, mark, structure):
"""
Mark upward-connected region on segment map starting at node pt
Parameters:
* mdata Masked data array
* map Array mapping out segments
* pt Index of starting node
* mark Integer to mark map with.
Returns nothing but modified mdata mask and map
"""
if mdata.mask[pt] == True:
return
else:
map[pt] = mark
mdata[pt] = ma.masked
Q = [pt]
while Q:
pt = Q.pop(0)
v = mdata.data[pt]
# Check all neightbors
for new_pt in neighbors(pt, mdata.shape, structure):
if mdata.mask[new_pt] == False and mdata[new_pt] > v:
Q.append(new_pt)
map[new_pt] = mark
mdata[new_pt] = ma.masked
return
def find_upward(data, pt, thres, diag=False):
"""
Find points upward-connected to pt in data.
Parameters:
* data 2D array of data
* pt Starting point, bottom of peak, tuple.
* thres Threshold, above this nodes are considered noise.
* diag True or False to include diagonal neighbors in connection.
Return: list of indicies of upward connected nodes.
"""
# build structure array for defining feature connections
ndim = data.ndim
if diag:
structure = ndimage.generate_binary_structure(ndim, ndim)
else:
structure = ndimage.generate_binary_structure(ndim, 1)
if type(pt) == int:
pt = (pt, )
pt = tuple(pt)
shape = data.shape
if data[pt] > thres: # check that the initial point is below threshold.
return []
Q = [pt] # queue
segment = [pt]
while Q: # loop until Q is empty
pt = Q.pop(0) # remove first element of queue
v = data[pt] # value at current node
for new_pt in neighbors(pt, shape, structure): # check all neightbors
if thres > data[new_pt] > v and new_pt not in segment:
Q.append(new_pt)
segment.append(new_pt)
return segment
def find_upward(data, pt, thres, diag=False):
"""
Find points upward-connected to pt in data.
Parameters:
* data 2D array of data
* pt Starting point, bottom of peak, tuple.
* thres Threshold, above this nodes are considered noise.
* diag True or False to include diagonal neighbors in connection.
Return: list of indicies of upward connected nodes.
"""
# build structure array for defining feature connections
ndim = data.ndim
if diag:
structure = ndimage.generate_binary_structure(ndim, ndim)
else:
structure = ndimage.generate_binary_structure(ndim, 1)
if type(pt) == int:
pt = (pt,)
pt = tuple(pt)
shape = data.shape
if data[pt] > thres: # check that the initial point is below threshold.
return []
Q = [pt] # queue
segment = [pt]
while Q: # loop until Q is empty
pt = Q.pop(0) # remove first element of queue
v = data[pt] # value at current node
for new_pt in neighbors(pt, shape, structure): # check all neightbors
if thres > data[new_pt] > v and new_pt not in segment:
Q.append(new_pt)
segment.append(new_pt)
return segment
