Question

我正在使用python中的计算机视觉来检测手写符号。我在单个字符的数据集上训练了cnn，但是现在我希望能够从图像中提取字符，以便对单个字符进行预测。做这个的最好方式是什么？我将使用的手写文本不会草书，而且字符之间会有明显的分隔。

Answer 1

在下面的代码段中，boxes变量具有图像中每个字符的尺寸。

import cv2
import pytesseract

file = '/content/Captchas/image22.jpg'

img = cv2.imread(file)
h, w, _ = img.shape

boxes = pytesseract.image_to_boxes(img)

for b in boxes.splitlines():
    b = b.split(' ')
    img = cv2.rectangle(img, (int(b[1]), h - int(b[2])), (int(b[3]), h - int(b[4])), (0, 255, 0), 2)

cv2_imshow(img)
print(boxes)

Answer 2

您可以使用查找轮廓并将其与框绑定。

image = cv2.imread("filename") 
image = cv2.fastNlMeansDenoisingColored(image,None,10,10,7,21)
gray = cv2.cvtColor(image,cv2.COLOR_BGR2GRAY)

res,thresh = cv2.threshold(gray,150,255,cv2.THRESH_BINARY_INV) #threshold 
kernel = cv2.getStructuringElement(cv2.MORPH_CROSS,(3,3)) 

 dilated = cv2.dilate(thresh,kernel,iterations = 5) 

 val,contours, hierarchy = 
            cv2.findContours(dilated,cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_NONE) 

 coord = []
 for contour in contours:  
      [x,y,w,h] = cv2.boundingRect(contour)   
      if h>300 and w>300:   
          continue   
      if h<40 or w<40:   
          continue  
      coord.append((x,y,w,h)) 

 coord.sort(key=lambda tup:tup[0]) # if the image has only one sentence sort in one axis

 count = 0
 for cor in coord:
        [x,y,w,h] = cor
        t = image[y:y+h,x:x+w,:]
        cv2.imwrite(str(count)+".png",t)
  print("number of char in image:", count)

python中的字符分割

2 个答案: