def mark_outliers(error_list, trim_stddev):
print("Marking outliers...")
sum = 0.0
count = len(error_list)
# numerically it is better to sum up a list of floatting point
# numbers from smallest to biggest (error_list is sorted from
# biggest to smallest)
for line in reversed(error_list):
sum += line[0]
# stats on error values
print(" computing stats...")
mean = sum / count
stddev_sum = 0.0
for line in error_list:
error = line[0]
stddev_sum += (mean - error) * (mean - error)
stddev = math.sqrt(stddev_sum / count)
print("mean = %.4f stddev = %.4f" % (mean, stddev))
# mark match items to delete
print(" marking outliers...")
mark_count = 0
for line in error_list:
# print "line:", line
if line[0] > mean + stddev * trim_stddev:
cull.mark_feature(matches_orig, line[1], line[2], line[0])
cull.mark_feature(matches_opt, line[1], line[2], line[0])
mark_count += 1
return mark_count
def split_image_features(index, matches):
# iterate through the match dictionary and mark any matches for
# the specified image for deletion
print("Marking feature matches for image:", index)
count = 0
new_matches = []
for i, match in enumerate(matches):
found_index = False
found_group = False
for j, p in enumerate(match[2:]):
if p[0] == index:
found_index = True
elif proj.image_list[p[0]].name in groups[args.group]:
found_group = True
# split match if possible
if found_index and found_group:
count += 1
new_match = [list(match[0]), -1]
for j, p in enumerate(match[2:]):
if proj.image_list[p[0]].name in groups[args.group]:
if p[0] != index:
cull.mark_feature(matches, i, j, 0)
# print('p:', p)
new_match.append(p)
if len(new_match) >= 4: # at least 2 images referenced
new_matches.append(new_match)
# add all the new match splits
for m in new_matches:
matches.append(m)
return count, len(new_matches)
