cv-analysis-service/vidocp/layout_detection.py
Matthias Bisping 3d4b924426 Pull request #4: Restructuring and renaming of module
Merge in RR/vidocp from poly_to_rects_segmentation to master

Squashed commit of the following:

commit 3dffe067ef0bb4796eab22007eb6970b29f47822
Author: Matthias Bisping <matthias.bisping@iqser.com>
Date:   Sat Feb 5 16:10:28 2022 +0100

    readme updated

commit 448517205259134a8427b48d86d0d5331b726487
Author: Matthias Bisping <matthias.bisping@iqser.com>
Date:   Sat Feb 5 16:09:35 2022 +0100

    restructured dirs

commit 058c2971631c71d520b1a94ea75e249f9234ad87
Author: Matthias Bisping <matthias.bisping@iqser.com>
Date:   Sat Feb 5 15:57:08 2022 +0100

    renaming

commit 4e64a3d07f1dad76775955639157ec7b60e6ad38
Author: Matthias Bisping <matthias.bisping@iqser.com>
Date:   Sat Feb 5 15:46:03 2022 +0100

    readme updated

commit 728bedb13a2769b4652fd674ef26988efebcc7dc
Author: Matthias Bisping <matthias.bisping@iqser.com>
Date:   Sat Feb 5 15:33:42 2022 +0100

    added DVC

commit e2d5594afd6683d8207007d3a85d178dd0a3e546
Author: Matthias Bisping <matthias.bisping@iqser.com>
Date:   Sat Feb 5 14:49:09 2022 +0100

    renaming
2022-02-05 16:14:24 +01:00

54 lines
1.5 KiB
Python

from itertools import count
import cv2
import numpy as np
import pdf2image
from matplotlib import pyplot as plt
import imutils
def find_layout_boxes(image: np.array):
gray_scale = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
blurred = cv2.GaussianBlur(gray_scale, (5, 5), 1)
thresh = cv2.threshold(blurred, 253, 255, cv2.THRESH_BINARY)[1]
img_bin = ~thresh
line_min_width = 10
kernel_h = np.ones((10, line_min_width), np.uint8)
kernel_v = np.ones((line_min_width, 10), np.uint8)
img_bin_h = cv2.morphologyEx(img_bin, cv2.MORPH_OPEN, kernel_h)
img_bin_v = cv2.morphologyEx(img_bin, cv2.MORPH_OPEN, kernel_v)
img_bin_final = img_bin_h | img_bin_v
contours = cv2.findContours(img_bin_final, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
contours = imutils.grab_contours(contours)
for c in contours:
peri = cv2.arcLength(c, True)
approx = cv2.approxPolyDP(c, 0.04 * peri, True)
yield cv2.boundingRect(approx)
def annotate_layout_boxes(image, rects):
for rect in rects:
(x, y, w, h) = rect
cv2.rectangle(image, (x, y), (x + w, y + h), (0, 255, 0), 2)
return image
def annotate_layout_in_pdf(pdf_path, page_index=1):
page = pdf2image.convert_from_path(pdf_path, first_page=page_index + 1, last_page=page_index + 1)[0]
page = np.array(page)
layout_boxes = find_layout_boxes(page)
page = annotate_layout_boxes(page, layout_boxes)
fig, ax = plt.subplots(1, 1)
fig.set_size_inches(20, 20)
ax.imshow(page)
plt.show()