How can i improve quality of this receipt

imageprocessing

asked 2020-03-23 05:50:09 -0600

kani
1 ●1 ●1

updated 2020-03-23 07:57:45 -0600

I have a problem recognizing data from selected area showed on receipt screenshot image description

For recognition i'm using Tesseract, and this red area is being skipped. What kind of techniques can i try to improve quality of an image?

Original image: https://imgur.com/a/yByTnsg

So far i tried this simple methods:

img = cv2.imread(image)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
median = cv2.GaussianBlur(gray, (5, 5), 0)
thresh = cv2.threshold(median, 127, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1]

edit retag flag offensive close merge delete

add a comment

1 answer

Sort by » oldest newest most voted

answered 2020-03-23 10:03:04 -0600

juanmamdp
31 ●1 ●1

Here you have a good example and explanation of dewarping: https://mzucker.github.io/2016/08/15/... There are plenty of examples of code for this kind of problem, but I like the one above.