#coding: utf-8
import sys
import cv2
import numpy as np
from ..utils import cv2ArgumentParser
from ..opencv import cvui, cv2Project
from ..opencv import draw_text_with_bg, findBiggestContour, reorder_contour, draw_bboxes_xywh
[docs]def cvPaperScanner(argv=sys.argv[1:]):
"""Paper Scanner using OpenCV.
Please see :meth:`cv2ArgumentParser <pycharmers.utils.argparse_utils.cv2ArgumentParser>` for arguments.
Note:
When you run from the command line, execute as follows::
$ cv-paper-scanner --cam 0 --radio-width 200
+--------------------------------------------+
| Sample |
+============================================+
| .. image:: _images/cli.cvPaperScanner.gif |
+--------------------------------------------+
"""
parser = cv2ArgumentParser(prog="cv-paper-scan", description="Paper Scanner", add_help=True)
args = parser.parse_args(argv)
project = cv2Project(args=args)
labels = ["Original", "Gray", "Canny Edge", "Contours", "Biggest Contour", "Warp Prespective", "Warp Gray", "Adaptive Threshold"]
states = [i==0 for i in range(len(labels))]
threshold1 = [100]
threshold2 = [200]
eta_counter = [0.1]
def func(frame, monitor, frame_width, frame_height, gui_x, **kwargs):
cvui.text(where=monitor, x=gui_x+20, y=30, text="[Document Scanner]")
idx = cvui.radiobox(where=monitor, x=gui_x, y=60, labels=labels, states=states)
cvui.text(where=monitor, x=gui_x+20, y=245, text="[Canny Edge]")
cvui.text(where=monitor, x=gui_x, y=270, text="* Low threshold")
cvui.text(where=monitor, x=gui_x, y=345, text="* High threshold")
cvui.text(where=monitor, x=gui_x+20, y=450, text="[Biggest Counter]")
cvui.text(where=monitor, x=gui_x, y=475, text="* eta")
th1 = cvui.trackbar(where=monitor, x=gui_x, y=290, width=150, value=threshold1, min=0., max=255.)
th2 = cvui.trackbar(where=monitor, x=gui_x, y=380, width=150, value=threshold2, min=0., max=255.)
eta = cvui.counter(where=monitor, x=gui_x+30, y=500, value=eta_counter, step=0.01, fmt="%.2f")
img_bgr = frame.copy()
# Gray
if idx>=1:
# Convert image to Gray scale.
frame = cv2.cvtColor(src=frame, code=cv2.COLOR_BGR2GRAY)
# Canny Edge
if idx>=2:
# Add Gaussian Blur.
img_blur = cv2.GaussianBlur(src=frame, ksize=(5, 5), sigmaX=1)
# APPLY Canny Blur.
img_th = cv2.Canny(image=img_blur, threshold1=th1, threshold2=th2)
# Apply Dilation & Erosion.
kernel = np.ones(shape=(5, 5), dtype=np.uint8)
frame = cv2.erode(src=cv2.dilate(src=img_th, kernel=kernel, iterations=2), kernel=kernel, iterations=1)
# Contours"
if idx>=3:
# Find All Contours.
contours, hierarchy = cv2.findContours(image=frame, mode=cv2.RETR_EXTERNAL, method=cv2.CHAIN_APPROX_SIMPLE)
img_binary = frame.copy()
frame = cv2.drawContours(image=frame, contours=contours, contourIdx=-1, color=(0, 255, 0), thickness=10)
# Biggest Contour
if idx>=4:
# Find the biggest Contour
biggest_contour, max_area = findBiggestContour(contours=contours, eta=eta)
if max_area == 0:
draw_text_with_bg(img=frame, text="Could not find the closed contours.", org=(10,50))
idx = 4
else:
# Draw the biggest contour
frame = img_binary
biggest_contour = reorder_contour(biggest_contour)
frame = cv2.drawContours(image=frame, contours=biggest_contour, contourIdx=-1, color=(0, 255, 0), thickness=20)
# Warp Prespective
if idx>=5:
matrix = cv2.getPerspectiveTransform(src=np.float32(biggest_contour), dst=np.float32([[0, 0],[frame_width, 0], [0, frame_height],[frame_width, frame_height]]))
frame = cv2.warpPerspective(src=img_bgr, M=matrix, dsize=(frame_width, frame_height))
# Warp Gray
if idx>=6:
frame = cv2.cvtColor(src=frame, code=cv2.COLOR_BGR2GRAY)
# Adaptive Threshold
if idx>=7:
frame = cv2.adaptiveThreshold(src=frame, maxValue=255, adaptiveMethod=1, thresholdType=1, blockSize=7, C=2)
frame = cv2.bitwise_not(src=frame)
frame = cv2.medianBlur(src=frame, ksize=3)
return frame
project.wrap(func=func)
