人民币字符识别——opencv和开源ocr模型两种方式识别

方法一：是用之前学习opencv时的一个课程——信用卡数字识别的思路来做的。课程的B站链接：https://www.bilibili.com/video/BV1oJ411D71z?p=9
方法二：ocr识别

样板字符与人民币图像准备

首先需要准备下样本字符，作为匹配的标准。我这里是直接在ppt中大概肉眼匹配了一组和人民币字符接近的字符，一共设置了A-Z，0-9，36个字符：

人民币图像采集，或者网上有人民币的数据集。

方法一：opencv

主要流程

1.myutils.py

该脚本为此次所用到的一些功能函数

import cv2
import os

def sort_contours(cnts, method="left-to-right"):
    reverse = False
    i = 0
    if method == "right-to-left" or method == "bottom-to-top":
        reverse = True
    if method == "top-to-bottom" or method == "bottom-to-top":
        i = 1
    boundingBoxes = [cv2.boundingRect(c) for c in cnts] #用一个最小的矩形，把找到的形状包起来x,y,h,w
    (cnts, boundingBoxes) = zip(*sorted(zip(cnts, boundingBoxes),
                                        key=lambda b: b[1][i], reverse=reverse))
    return cnts, boundingBoxes

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

def cv_show(name,img):
    cv2.imshow(name, img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()

def img_save(file_name,img):
    path='./result/'
    if not os.path.exists(path):
        os.mkdir(path)
    file=path+file_name
    cv2.imwrite(file,img)

2.rmb.py

from imutils import contours
import numpy as np
import os
import argparse
import cv2
import myutils
from myutils import cv_show,img_save

# 设置参数
ap = argparse.ArgumentParser()
ap.add_argument("-i", "--image", required=True,
   help="path to input image")
ap.add_argument("-t", "--template", required=True,
   help="path to template OCR-A image")
args = vars(ap.parse_args())

# 读取一个模板图像
img = cv2.imread(args["template"])
cv_show('img', img)
img_save('1.png', img)
# 灰度图
ref = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
cv_show('ref', ref)
img_save('2.png', ref)
# 二值图像
ref = cv2.threshold(ref, 10, 255, cv2.THRESH_BINARY_INV)[1]
cv_show('ref', ref)
img_save('3.png', ref)

# 计算轮廓
refCnts, hierarchy = cv2.findContours(ref.copy(), cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)

cv2.drawContours(img,refCnts,-1,(0,0,255),3)
cv_show('img',img)
img_save('4.png',img)
print(np.array(refCnts).shape)
refCnts = myutils.sort_contours(refCnts, method="left-to-right")[0] #排序，从左到右，从上到下
digits = {}

list=['A','B','C','D','E','F','G','H','I','J','K','L','M','N','O','P','Q','R','S','T',\
     'U','V','W','X','Y','Z','0','1','2','3','4','5','6','7','8','9']
# 遍历每一个轮廓
for (i, c) in enumerate(refCnts):
   # print(i)
   # 计算外接矩形并且resize成合适大小
   (x, y, w, h) = cv2.boundingRect(c)
   roi = ref[y:y + h, x:x + w]
   roi = cv2.resize(roi, (57, 88))
   # 每一个数字对应每一个模板
   digits[list[i]] = roi

# for key,value in digits.items():
#  print(key)


# 初始化卷积核
rectKernel = cv2.getStructuringElement(cv2.MORPH_RECT, (12, 3))
sqKernel = cv2.getStructuringElement(cv2.MORPH_RECT, (7, 7))
#读取输入图像，预处理
image = cv2.imread(args["image"])
cv_show('image',image)
img_save('5.png',image)
print(image.shape)
# image = myutils.resize(image, width=300)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
cv_show('gray',gray)
img_save('6.png',gray)
#礼帽操作，突出更明亮的区域
tophat = cv2.morphologyEx(gray, cv2.MORPH_TOPHAT, rectKernel)
cv_show('tophat',tophat)
img_save('7.png',tophat)

gradX = cv2.Sobel(tophat, ddepth=cv2.CV_32F, dx=1, dy=0, #ksize=-1相当于用3*3的
   ksize=-1)
gradX = np.absolute(gradX)
(minVal, maxVal) = (np.min(gradX), np.max(gradX))
gradX = (255 * ((gradX - minVal) / (maxVal - minVal)))
gradX = gradX.astype("uint8")
print (np.array(gradX).shape)
cv_show('gradX',gradX)
img_save('8.png',gradX)

#通过闭操作（先膨胀，再腐蚀）将数字连在一起
gradX = cv2.morphologyEx(gradX, cv2.MORPH_CLOSE, rectKernel)
cv_show('gradX',gradX)
img_save('9.png',gradX)
#THRESH_OTSU会自动寻找合适的阈值，适合双峰，需把阈值参数设置为0
thresh = cv2.threshold(gradX, 0, 255,
   cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1]
cv_show('thresh',thresh)
img_save('10.png',thresh)

#再来一个闭操作
thresh = cv2.morphologyEx(thresh, cv2.MORPH_CLOSE, sqKernel) #再来一个闭操作
cv_show('thresh',thresh)
img_save('11.png',thresh)


# 计算轮廓
threshCnts, hierarchy = cv2.findContours(thresh.copy(), cv2.RETR_EXTERNAL,
   cv2.CHAIN_APPROX_SIMPLE)

cnts = threshCnts
cur_img = image.copy()
cv2.drawContours(cur_img,cnts,-1,(0,0,255),3)
cv_show('img',cur_img)
img_save('12.png',cur_img)
locs = []  # 用于存储目标轮廓区域的（xywh）

# 遍历轮廓
for (i, c) in enumerate(cnts):
   # 计算矩形
   (x, y, w, h) = cv2.boundingRect(c)
   # 通过长宽比进行筛选需要识别的数字区域
   ar = w / float(h)

   # 选择合适的区域，根据实际任务来，这里的基本都是四个数字一组
   if ar > 4 and ar < 8:

      if (w > 130 and w <160 ) and (h > 20 and h < 30):
         #符合的留下来
         locs.append((x, y, w, h))

print(locs)

# 将符合的轮廓从左到右排序
# locs = sorted(locs, key=lambda x:x[0])   #这里只有一个符合的，不用排序
output = []

# 遍历轮廓中的数字
for (i, (gX, gY, gW, gH)) in enumerate(locs):
   # initialize the list of group digits
   print("x,y,w,h",gX,gY,gW,gH)
   groupOutput = []

   # 根据坐标提取每一个组
   group = gray[gY - 5:gY + gH + 5, gX - 5:gX + gW + 5]  
   cv_show('group',group)
   img_save('13_'+str(gX)+'.png', group)
   # 预处理
   group = cv2.threshold(group, 0, 255,
      cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1]
   cv_show('group',group)
   img_save('14_'+str(gX)+'.png', group)
   #做个反色
   group=cv2.bitwise_not(group)

   # 计算每一组的轮廓
   digitCnts,hierarchy = cv2.findContours(group.copy(), cv2.RETR_EXTERNAL,
      cv2.CHAIN_APPROX_SIMPLE)
   digitCnts = contours.sort_contours(digitCnts,
      method="left-to-right")[0]

   # 计算每一组中的每一个数值
   a=1
   for c in digitCnts:
      # 找到当前数值的轮廓，resize成合适的的大小
      (x, y, w, h) = cv2.boundingRect(c)
      print(x,y,w,h)
      roi = group[y:y + h, x:x + w]
      roi = cv2.resize(roi, (57, 88))
      cv_show('roi',roi)
      img_save('15_' +str(a) + '.png', roi)
      a+=1

      # 计算匹配得分
      scores = []

      # 在模板中计算每一个得分
      for (digit, digitROI) in digits.items():
         # 模板匹配
         result = cv2.matchTemplate(roi, digitROI,
            cv2.TM_CCOEFF)
         (_, score, _, _) = cv2.minMaxLoc(result)
         scores.append(score)

      # 得到最合适的数字
      groupOutput.append(str(np.argmax(scores)))
   print(groupOutput)
   #将对应位置的结果索引出来
   result=[]
   for i in groupOutput:
      result_sor=list[int(i)]
      result.append(result_sor)
   print('result',result)

   # 画出来
   cv2.rectangle(image, (gX - 5, gY - 5),
      (gX + gW + 5, gY + gH + 5), (0, 0, 255), 1)
   cv2.putText(image, "".join(result), (gX, gY - 15),
      cv2.FONT_HERSHEY_SIMPLEX, 0.65, (0, 0, 255), 2)

   # 得到结果
   output.extend(groupOutput)

cv2.imshow("Image", image)
img_save("16.png", image)
cv2.waitKey(0)

结果显示

方法二：easyocr库使用

easyocr官方github.具体的说明文档见github.
使用效果：

支持的识别语言：

代码

import easyocr
import cv2
import numpy as np
from PIL import Image, ImageDraw, ImageFont
import os

fontC = ImageFont.truetype("platech.ttf", 23, 0)
def cv2ImgAddText(img, text, left, top, textColor=(255, 0, 0), textSize=30):
    if (isinstance(img, np.ndarray)):  # 判断是否OpenCV图片类型
        img = Image.fromarray(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))
    # 创建一个可以在给定图像上绘图的对象
    draw = ImageDraw.Draw(img)
    # 字体的格式
    fontStyle = ImageFont.truetype(
        "font/simsun.ttc", textSize, encoding="utf-8")
    # 绘制文本
    draw.text((left, top), text, textColor, font=fontC)
    # 转换回OpenCV格式
    return cv2.cvtColor(np.asarray(img), cv2.COLOR_RGB2BGR)

def image_detect():
    img=cv2.imread('./outdoorAA-1.jpg')
    reader = easyocr.Reader(['ch_sim','en'])
    result = reader.readtext(img)
    for a,b,c in result:
        print(a,b,c)
        pts=np.array(a,np.int32)
        pts=pts.reshape((-1,1,2))
        left=a[0][0]
        top=a[0][1]
        if len(b)>0:
            cv2.polylines(img,[pts],True,(0,0,255),thickness=2)
            img=cv2ImgAddText(img,str(b)+'-'+str(round(c,3)),left,top-30)

    cv2.imshow('img',img)
    cv2.imwrite('./1.jpg',img)
    cv2.waitKey(0)

def dir_detect():
    test_dir = "./public_test_data"  # 图片读入路径
    # fw = open("./test-results/results.txt", 'w+')  # 以覆盖写方式打开文件，如果不存在，则新建一个
    fw= "./test-results/results.txt"
    os.makedirs(os.path.dirname(fw), exist_ok=True)
    fw = open("./test-results/No14007mresults.txt", 'w+')
    for f in os.listdir(test_dir):
        try:
            if f.endswith(".jpg") or f.endswith("JPG") or f.endswith("png"):
                # print("---------"+f+"----------------");
                cpp = test_dir + "/" + f    # 生成完整路径
                img=cv2.imread(cpp)
                reader = easyocr.Reader(['ch_sim', 'en'])
                result = reader.readtext(img)
                for a,b, c in result:
                    # print(a, b, c)
                    pts = np.array(a, np.int32)
                    pts = pts.reshape((-1, 1, 2))
                    left = a[0][0]
                    top = a[0][1]
                    if (len(b) == 10 )and (' 'not in b): # 这个是将检测到的结果筛选
                        print(b)
                        cv2.polylines(img, [pts], True, (0, 0, 255), thickness=2)
                        img = cv2ImgAddText(img, str(b) + '-' + str(round(c, 3)), left, top - 30)
                        info=str(f)+'\t'+str(b)+'\t'+str(round(c,3))
                        fw.write(info)
                cv2.imwrite("./test-results/" + f, img)
        except Exception as e:
            print(e)  # 输出异常信息，调试用，发布时应注释掉
            continue
    fw.close()
    cv2.destroyAllWindows()

# 两种检测方式，一个是检测单张，一个是检测整个文件夹。
# dir_detect()
image_detect()

效果