9d98945ff9
Merge in RR/cv-analysis from new_pyinfra to master
Squashed commit of the following:
commit f7a01a90aad1c402ac537de5bdf15df628ad54df
Author: Julius Unverfehrt <julius.unverfehrt@iqser.com>
Date: Wed Jul 27 10:40:59 2022 +0200
fix typo
commit ff4d549fac5b612c2d391ae85823c5eca1e91916
Author: Julius Unverfehrt <julius.unverfehrt@iqser.com>
Date: Wed Jul 27 10:34:04 2022 +0200
adjust build scripts for new pyinfra
commit ecd70f60d46406d8b6cc7f36a1533d706c917ca8
Author: Julius Unverfehrt <julius.unverfehrt@iqser.com>
Date: Wed Jul 27 09:42:55 2022 +0200
simplify logging by using default configurations
commit 20193c14c940eed2b0a7a72058167e26064119d0
Author: Julius Unverfehrt <julius.unverfehrt@iqser.com>
Date: Tue Jul 26 17:16:57 2022 +0200
tidy-up, refactor config logic to not dependent on external files
commit d8069cd4d404a570bb04a04278161669d1c83332
Author: Isaac Riley <Isaac.Riley@iqser.com>
Date: Tue Jul 26 15:14:59 2022 +0200
update pyinfra
commit c3bc11037cca9baf016043ab997c566f5b4a2586
Author: Isaac Riley <Isaac.Riley@iqser.com>
Date: Tue Jul 26 15:09:14 2022 +0200
repair tests
commit 6f4e4f2863ee16ae056c1d432f663858c5f10221
Author: Isaac Riley <Isaac.Riley@iqser.com>
Date: Tue Jul 26 14:52:38 2022 +0200
updated server logic to work with new pyinfra; update scripts for pyinfra as submodule
commit 2a18dba81de5ee84d0bdf0e77f478693e8d8aef4
Author: Isaac Riley <Isaac.Riley@iqser.com>
Date: Tue Jul 26 14:10:41 2022 +0200
formatting
commit d87ce9328de9aa2341228af9b24473d5e583504e
Author: Isaac Riley <Isaac.Riley@iqser.com>
Date: Tue Jul 26 14:10:11 2022 +0200
make server logic compatible with new pyinfra
42 lines
1.3 KiB
Python
42 lines
1.3 KiB
Python
from functools import partial
|
|
|
|
import cv2
|
|
import numpy as np
|
|
from iteration_utilities import starfilter, first
|
|
|
|
from cv_analysis.utils.filters import is_large_enough, is_filled, is_boxy
|
|
from cv_analysis.utils.visual_logging import vizlogger
|
|
|
|
|
|
def is_likely_redaction(contour, hierarchy, min_area):
|
|
return is_filled(hierarchy) and is_boxy(contour) and is_large_enough(contour, min_area)
|
|
|
|
|
|
def find_redactions(image: np.array, min_normalized_area=200000):
|
|
vizlogger.debug(image, "redactions01_start.png")
|
|
min_normalized_area /= 200 # Assumes 200 DPI PDF -> image conversion resolution
|
|
|
|
if len(image.shape) > 2:
|
|
gray = ~cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
|
|
else:
|
|
gray = ~image
|
|
vizlogger.debug(gray, "redactions02_gray.png")
|
|
blurred = cv2.GaussianBlur(gray, (5, 5), 1)
|
|
vizlogger.debug(blurred, "redactions03_blur.png")
|
|
thresh = cv2.threshold(blurred, 252, 255, cv2.THRESH_BINARY)[1]
|
|
vizlogger.debug(blurred, "redactions04_threshold.png")
|
|
|
|
contours, hierarchies = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_NONE)
|
|
|
|
try:
|
|
contours = map(
|
|
first,
|
|
starfilter(
|
|
partial(is_likely_redaction, min_area=min_normalized_area),
|
|
zip(contours, hierarchies[0]),
|
|
),
|
|
)
|
|
return list(contours)
|
|
except:
|
|
return []
|