I'm trying to draw a filled rectangle around some OCR'd text.
d = pytesseract.image_to_data(image, output_type=Output.DICT)
os.remove(filename);
n_boxes = len(d['level'])
for i in range(n_boxes):
(x, y, w, h) = (d['left'][i], d['top'][i], d['width'][i], d['height'][i])
cv2.rectangle(image, (x, y), (x + w, y + h), (0, 0, 0), 1)
cv2.imwrite("./output/{}.png".format("blurred_" + str(os.getpid())), image);
This line seems to be the problem:
cv2.rectangle(image, (x, y), (x + w, y + h), (0, 0, 0), 1)
When I replace 1 with -1, which should draw the filled rectangle, it instead fills the entire image with the filled rectangle.
Without fill, you can see what should be colored:
Here's the full code, if needed:
from PIL import Image
import pytesseract
from pytesseract import Output
import argparse, cv2, os
import numpy as np
ap = argparse.ArgumentParser()
ap.add_argument("-i", "--image", required=True,
help="path of image to be blurrified")
ap.add_argument("-b", "--blurtype", required=True,
help="blur [words] or [character]s")
ap.add_argument("-p", "--preprocess", type=str, default="thresh",
help="preprocess type")
args = vars(ap.parse_args())
image = cv2.imread(args["image"])
h, w, _ = image.shape
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
if args["preprocess"] == "thresh":
gray = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1]
elif args["preprocess"] == "blur":
gray = cv2.medianBlur(gray, 3)
filename = "./temp/{}.png".format("grayscale_" + str(os.getpid()))
cv2.imwrite(filename, gray)
if args["blurtype"] == "character":
text = pytesseract.image_to_boxes(Image.open(filename))
os.remove(filename);
for b in text.splitlines():
b = b.split(' ')
image = cv2.rectangle(image, (int(b[1]), h - int(b[2])), (int(b[3]), h - int(b[4])), (0, 0, 0), -1)
cv2.imwrite("./output/{}.png".format("blurred_" + str(os.getpid())), image);
elif args["blurtype"]== "words":
d = pytesseract.image_to_data(image, output_type=Output.DICT)
os.remove(filename);
n_boxes = len(d['level'])
for i in range(n_boxes):
(x, y, w, h) = (d['left'][i], d['top'][i], d['width'][i], d['height'][i])
cv2.rectangle(image, (x, y), (x + w, y + h), (0, 0, 0), 1)
cv2.imwrite("./output/{}.png".format("blurred_" + str(os.getpid())), image);
Help would be greatly appreciated!
The pytesseract is making a bounding box of the complete image also. To view this, with thickness 1, making rectangle of (0, 255, 0) color. Thus, when you pass thickness = -1, the image's bounding box is also colored and hence the compete image is being colored. To avoid this, reject that rectangle.