How to Recognize Credit Card Numbers with OpenCV: A Step‑by‑Step Tutorial

Practice Is the Sole Test of Truth

Because learning OpenCV in a rigid way feels dull, I followed a project‑oriented tutorial and started coding.

1. Case Introduction

Provide a digit template on a credit card.

Requirement: Recognize the digits on the credit card, print them on the original image, and optionally save them as text. The approach is similar to license‑plate recognition.

Example

Original image

Processed image

Overall Steps

1. Load template 2. Preprocess template (grayscale, threshold, contour detection, sorting) 3. Preprocess input image (resize, grayscale, tophat, Sobel, morphological operations, threshold) 4. Match digits with template.

1. Load Template and Import Packages

import cv2 as cv
import numpy as np
import myutils

def cv_show(name, img):
    cv.imshow(name, img)
    cv.waitKey(0)
# Load template image
img = cv.imread("images/ocr_a_reference.png")

2. Template Preprocessing

Convert to grayscale, invert binary, find contours, draw them.

# Convert to gray
ref = cv.cvtColor(img, cv.COLOR_BGR2GRAY)
# Invert binary
ref = cv.threshold(ref, 10, 255, cv.THRESH_BINARY_INV)[1]
# Find contours
refCnts, hierarchy = cv.findContours(ref.copy(), cv.RETR_EXTERNAL, cv.CHAIN_APPROX_NONE)
cv.drawContours(img, refCnts, -1, (0, 0, 255), 2)

Sort contours left‑to‑right and store each digit ROI in a dictionary.

refCnts = myutils.sort_contours(refCnts)[0]
digits = {}
for (i, c) in enumerate(refCnts):
    (x, y, w, h) = cv.boundingRect(c)
    roi = ref[y:y+h, x:x+w]
    roi = cv.resize(roi, (57, 88))
    digits[i] = roi

3. Input Image Preprocessing

Initialize structuring elements, read and resize the credit‑card image, convert to gray.

rectKernel = cv.getStructuringElement(cv.MORPH_RECT, (9, 3))
sqKernel = cv.getStructuringElement(cv.MORPH_RECT, (5, 5))
card_image = cv.imread("images/credit_card_01.png")
card_image = myutils.resize(card_image, width=300)
gray = cv.cvtColor(card_image, cv.COLOR_BGR2GRAY)

Apply tophat, Sobel gradient, morphological closing, and Otsu threshold.

tophat = cv.morphologyEx(gray, cv.MORPH_TOPHAT, rectKernel)
gradX = cv.Sobel(tophat, cv.CV_32F, 1, 0, ksize=3)
gradX = np.absolute(gradX)
min_Val, max_val = np.min(gradX), np.max(gradX)
gradX = (255 * (gradX - min_Val) / (max_val - min_Val)).astype("uint8")
gradX = cv.morphologyEx(gradX, cv.MORPH_CLOSE, rectKernel)
thresh = cv.threshold(gradX, 0, 255, cv.THRESH_BINARY | cv.THRESH_OTSU)[1]
thresh = cv.morphologyEx(thresh, cv.MORPH_CLOSE, sqKernel)

Find contours of the digit groups.

threshCnts = cv.findContours(thresh.copy(), cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE)[0]
card_copy = card_image.copy()
cv.drawContours(card_copy, threshCnts, -1, (0, 0, 255), 2)

4. Template Matching

Select contour bounding boxes with appropriate aspect ratio, sort them left‑to‑right.

locs = []
for i, c in enumerate(threshCnts):
    x, y, w, h = cv.boundingRect(c)
    ar = w / float(h)
    if 2.5 < ar < 4.0 and 40 < w < 55 and 10 < h < 20:
        locs.append((x, y, w, h))
locs = sorted(locs, key=lambda x: x[0])

For each group, extract the region, threshold, find digit contours, sort them, and match each digit against the template using cv.matchTemplate.

output = []
for (gx, gy, gw, gh) in locs:
    group = gray[gy-5:gy+gh+5, gx-5:gx+gw+5]
    group = cv.threshold(group, 0, 255, cv.THRESH_BINARY | cv.THRESH_OTSU)[1]
    digitCnts = cv.findContours(group.copy(), cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE)[0]
    digitCnts = myutils.sort_contours(digitCnts, method="left-to-right")[0]
    groupOutput = []
    for c in digitCnts:
        (x, y, w, h) = cv.boundingRect(c)
        roi = group[y:y+h, x:x+w]
        roi = cv.resize(roi, (57, 88))
        scores = []
        for (digit, digitROI) in digits.items():
            res = cv.matchTemplate(roi, digitROI, cv.TM_CCOEFF)
            scores.append(cv.minMaxLoc(res)[1])
        groupOutput.append(str(np.argmax(scores)))
    cv.rectangle(card_image, (gx-5, gy-5), (gx+gw+5, gy+gh+5), (0,0,255), 1)
    cv.putText(card_image, "".join(groupOutput), (gx, gy-15), cv.FONT_HERSHEY_SIMPLEX, 0.65, (0,0,255), 2)

Display the final image with recognized digits.

cv.imshow("Output_image", card_image)
cv.waitKey(0)

Conclusion

This credit‑card recognition example demonstrates basic image‑processing operations and provides a friendly entry point for newcomers to OpenCV.