Opencv learning notes 5 -- document scanning +ocr character recognition

File scanning

#  Import toolkit 
import numpy as np
import argparse
import cv2
import matplotlib.pyplot as plt#Matplotlib yes RGB

Defined function

#  Graphic display 
def cv_show(name,img):
    b,g,r = cv2.split(img)
    img_rgb = cv2.merge((r,g,b))
    plt.imshow(img_rgb)
    plt.show()
def cv_show1(name,img):
    plt.imshow(img)
    plt.show()
    cv2.imshow(name,img)
    cv2.waitKey()
    cv2.destroyAllWindows()

def order_points(pts):
    	#  altogether 4 Coordinates 
	rect = np.zeros((4, 2), dtype = "float32")

	#  Find the corresponding coordinates in order 0123 Namely   Top left , The upper right , The lower right , The lower left 
	#  Calculate top left , The lower right 
	s = pts.sum(axis = 1)# Addition of horizontal and vertical coordinates , The biggest one is the bottom right , The smallest is the upper left 
	rect[0] = pts[np.argmin(s)]
	rect[2] = pts[np.argmax(s)]

	#  Count right up and left down 
	diff = np.diff(pts, axis = 1)
	rect[1] = pts[np.argmin(diff)]
	rect[3] = pts[np.argmax(diff)]

	return rect

def four_point_transform(image, pts):
    	#  Get the input coordinate point 
	rect = order_points(pts)#pts Is the coordinates of the four points on the original drawing 
	(tl, tr, br, bl) = rect

	#  Calculate the input w and h value 
	widthA = np.sqrt(((br[0] - bl[0]) ** 2) + ((br[1] - bl[1]) ** 2))
	widthB = np.sqrt(((tr[0] - tl[0]) ** 2) + ((tr[1] - tl[1]) ** 2))
	maxWidth = max(int(widthA), int(widthB))

	heightA = np.sqrt(((tr[0] - br[0]) ** 2) + ((tr[1] - br[1]) ** 2))
	heightB = np.sqrt(((tl[0] - bl[0]) ** 2) + ((tl[1] - bl[1]) ** 2))
	maxHeight = max(int(heightA), int(heightB))

	#  Corresponding coordinate position after transformation 
	dst = np.array([
		[0, 0],
		[maxWidth - 1, 0],
		[maxWidth - 1, maxHeight - 1],
		[0, maxHeight - 1]], dtype = "float32")

	#  Calculate the transformation matrix 
	M = cv2.getPerspectiveTransform(rect, dst)# from rect To dst The transformation matrix of 
	warped = cv2.warpPerspective(image, M, (maxWidth, maxHeight))# Get the transformation result 

	#  Return the result after transformation 
	return warped

def resize(image, width=None, height=None, inter=cv2.INTER_AREA):
	dim = None
	(h, w) = image.shape[:2]
	if width is None and height is None:
		return image
	if width is None:
		r = height / float(h)
		dim = (int(w * r), height)
	else:
		r = width / float(w)
		dim = (width, int(h * r))
	resized = cv2.resize(image, dim, interpolation=inter)
	return resized

#  Read input 
image = cv2.imread("./images/receipt.jpg")
# The coordinates will change the same 
cv_show("Image",image)
ratio = image.shape[0] / 500.0
print(image.shape[0])
orig = image.copy()

2448

image = resize(orig, height = 500)
cv_show("Image",image)

edge detection

#  Preprocessing 
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
gray = cv2.GaussianBlur(gray, (5, 5), 0)
edged = cv2.Canny(gray, 75, 200)

#  Show the preprocessing results 
print("STEP 1:  edge detection ")
cv_show("Image", image)
cv_show1("Edged", edged)

STEP 1:  edge detection 

#  Contour detection 
cnts = cv2.findContours(edged.copy(), cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)[1]# Detect contour 
cnts = sorted(cnts, key = cv2.contourArea, reverse = True)[:5]# Sort the outline according to the area 
#cv2.drawContours(image, cnts, -1, (0, 255, 0), 2)
#cv_show("Outline", image)

#  Traverse the outline 
for c in cnts:
	#  Calculate the contour approximation 
	peri = cv2.arcLength(c, True)# Calculate contour perimeter 
	# C Represents the set of input points 
	# epsilon Represents the maximum distance from the original contour to the approximate contour , It's an accuracy parameter 
	# True It means closed 
	approx = cv2.approxPolyDP(c, 0.02 * peri, True)

	# 4 Take it out at one point 
	if len(approx) == 4:
		screenCnt = approx
		break

Get profile

#  Display the results 
print("STEP 2:  Get profile ")
cv2.drawContours(image, [screenCnt], -1, (0, 255, 0), 2)
cv_show("Outline", image)

STEP 2:  Get profile 

#  Perspective transformation 
warped = four_point_transform(orig, screenCnt.reshape(4, 2) * ratio)
print(screenCnt.reshape(4, 2))
print(screenCnt.reshape(4, 2).sum(axis = 1))

[[465 110]
 [113 137]
 [147 375]
 [474 323]]
[575 250 522 797]

Transformation

#  Binary processing 
warped = cv2.cvtColor(warped, cv2.COLOR_BGR2GRAY)
ref = cv2.threshold(warped, 100, 255, cv2.THRESH_BINARY)[1]
cv2.imwrite('scan.jpg', ref)
#  Display the results 
print("STEP 3:  Transformation ")
cv_show("Original", resize(orig, height = 650))
cv_show1("Scanned", resize(ref, height = 650))

STEP 3:  Transformation 

OCR Character recognition

Environment configuration

install tesseract-ocr-w64-setup-v5.0.1.20220118.exe

1. https://digi.bib.uni-mannheim.de/tesseract/
2. Configure environment variables such as E:\Program Files (x86)\Tesseract-OCR
3. tesseract -v To test
4. tesseract XXX.png Get the results
5. pip install pytesseract
6. anaconda lib site-packges pytesseract pytesseract.py
7. tesseract_cmd Change to absolute path

Code

from PIL import Image
import pytesseract
import cv2
import os
import matplotlib.pyplot as plt#Matplotlib yes RGB

#  Graphic display 
def cv_show(name,img):
    b,g,r = cv2.split(img)
    img_rgb = cv2.merge((r,g,b))
    plt.imshow(img_rgb)
    plt.show()
def cv_show1(name,img):
    plt.imshow(img)
    plt.show()
    cv2.imshow(name,img)
    cv2.waitKey()
    cv2.destroyAllWindows()

preprocess = 'thresh' #thresh

image = cv2.imread('scan.jpg')
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

if preprocess == "thresh":
    gray = cv2.threshold(gray, 0, 255,cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1]# Adaptive binarization 

if preprocess == "blur":
    gray = cv2.medianBlur(gray, 3)# Median ambiguity 

filename = "{}.png".format(os.getpid())
cv2.imwrite(filename, gray)

True

text = pytesseract.image_to_string(Image.open(filename))
print(text)
os.remove(filename)

cv_show("Image", image)
cv_show1("Output", gray)

we KK Re KK KK OK OK KK

WHOLE FOODS MARKET - WESTPORT, CT 06880
399 POST RD WEST - [phone]

365
365
365
365

BROTH

BACON
BACON
BACON
BACUN

LS
LS
LS
LS

CHIC

FLOUR ALMUND
CHKN BRST BNLSS SK
HEAVY CREAM
BALSMC REDUCT
GRND 85/15

BEEF

JUICE

COF CASHEW

L

DOCS PINT ORGANIC
HNY ALMOND BUTTER

xeene TAX

.00

BAL

NP
NP
NP
NP
NP
NP
NP
NP
NP
NP
NP
NP
NP

4
4
4

99

.99
.99

mal

7 7 T

mana Ramm