cv2条码识别

一、前言

本文主要使用opencv 的Python接口cv2，原理较为简单，单张先通过二值化找轮廓的方式找到轮廓的四个点，然后对点做排序，之后再计算轮廓的外接矩形的4个点，由这2组点做 warpPerspective 透射变换得到矫正的条码，最后使用zbar库做识别。

二、代码

import os,cv2,time
import numpy as np

import math  
from math import *

from scipy import misc, ndimage
import matplotlib.pyplot as plt

import xml.etree.ElementTree as ET

def getBoxFromCnt(cnt):
    x_coord = cnt[:, :, 0]
    y_coord = cnt[:, :, 1]
            
    xmin = int(np.min(x_coord))
    xmax = int(np.max(x_coord))
    ymin = int(np.min(y_coord))
    ymax = int(np.max(y_coord))
    
    return xmin,xmax,ymin,ymax

def getCnt(bar_img):
    gray_roi = cv2.cvtColor(bar_img , cv2.COLOR_BGR2GRAY)
    blurimg = cv2.GaussianBlur(gray_roi, (9, 9), 0)  #这个数据有点固定死，条件要求高
    ret, th = cv2.threshold(blurimg, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)
    kernel = np.ones((9,9),np.uint8)
    th_roi = cv2.dilate(th,kernel,iterations =1)

    Cannyimg = cv2.Canny(th_roi, 35, 189)  #这个数据有点固定死
   
    #显示边沿线条的位置
    lines = cv2.HoughLinesP(Cannyimg, 1, np.pi / 180, threshold=10, minLineLength=20, maxLineGap=10)
    lines = np.squeeze(lines)

    cnts = cv2.findContours(Cannyimg.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
    cnts = cnts[0]  
    pts,bar = [], [] 
    """
    cv2.namedWindow("th_roi" ,2)
    cv2.imshow("th_roi" , th_roi   )
    cv2.waitKey()    
    cv2.destroyAllWindows()     """
    max_area = 0
    if len(cnts) > 0:
        for cnt in cnts:
            xmin,xmax,ymin,ymax = getBoxFromCnt(cnt)            
            area = (xmax-xmin)*(ymax-ymin)
            if  area > max_area: 
                max_area = area
                max_cnt =  cnt        
                
        rect = cv2.minAreaRect(max_cnt)
        box = cv2.boxPoints(rect)
        box = np.int0(box)  # 获得矩形角点
        area = cv2.contourArea(box)
        width = rect[1][0]
        height = rect[1][1]
        """
        img2 = bar_img.copy()
        cv2.polylines(img2, [box], True, (0, 255, 0), 3)    
        cv2.namedWindow("img2" ,2)
        cv2.imshow("img2" ,  img2 )
        cv2.waitKey( )    
        cv2.destroyAllWindows()   """      
            
    #cv2.circle(bar_img,(pts[0][0],pts[0][1]),10,(255 ,0,0))
    if len(box) != 0: 
        pts = box
        minx = np.min(pts[:,0])
        maxx = np.max(pts[:,0])
        miny = np.min(pts[:,1])
        maxy = np.max(pts[:,1])
        bar = [minx,miny,maxx,maxy]
    
    return pts,bar #失败则返回[]

def sortByX(elem):
    return int(elem[0])
def sortByY(elem):
    return int(elem[1])

def orderPts(pts_np):
    dst_pts = []
    pts_1 = []
    pts_list = pts_np.tolist()
    pt_0 = pts_list.pop()
    pts_1.append(pt_0)
    least_dist = 10e8
    least_idx = 0
    for idx, pt in enumerate(pts_list):
        dist = np.sqrt( ((pt_0[0]-pt[0])**2) +( (pt_0[1]-pt[1])**2) )
        if dist<least_dist:
            least_idx = idx
            least_dist = dist
    
    pts_1.append(pts_list.pop(least_idx))
    pts_1.sort(key=sortByX,reverse=False)    
    pts_list.sort(key=sortByX,reverse=False) 
    
    if pts_1[0][0] < pts_list[0][0]:
         pts_1.sort(key=sortByY,reverse=False)
         pts_list.sort(key=sortByY,reverse=False)
         dst_pts = [pts_1[0], pts_list[0],pts_list[1] ,pts_1[1]  ]
    else:
         pts_1.sort(key=sortByY,reverse=False)
         pts_list.sort(key=sortByY,reverse=False)
         dst_pts =  [ pts_list[0], pts_1[0],pts_1[1] ,pts_list[1] ]
         
    degree = np.arctan(((dst_pts[1][1]-dst_pts[0][1] ) /(dst_pts[1][0]-dst_pts[0][0] +1e-7)))*180/np.pi #原来只是180度
     
    return np.array( dst_pts, dtype = "float32" ) 


def four_point_transform(image, pts):
    # obtain a consistent order of the points and unpack them
    # individually
    
    (tl, tr, br, bl) = pts
    
    # compute the width of the new image, which will be the
    # maximum distance between bottom-right and bottom-left
    # x-coordiates or the top-right and top-left x-coordinates
    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))
 
    # compute the height of the new image, which will be the
    # maximum distance between the top-right and bottom-right
    # y-coordinates or the top-left and bottom-left y-coordinates
    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))
 
    # now that we have the dimensions of the new image, construct
    # the set of destination points to obtain a "birds eye view",
    # (i.e. top-down view) of the image, again specifying points
    # in the top-left, top-right, bottom-right, and bottom-left
    # order
    dst = np.array([
        [0, 0],
        [maxWidth - 1, 0],
        [maxWidth - 1, maxHeight - 1],
        [0, maxHeight - 1]], dtype = "float32")
 
    # compute the perspective transform matrix and then apply it
    M = cv2.getPerspectiveTransform(pts, dst)
    warped = cv2.warpPerspective(image, M, (maxWidth, maxHeight))

    # return the warped image
    return warped

 
def recogBarcode(barcode_img):
    #调用cmd命令，获取cmd的输出
    bar_save_path = "temp.png" 
    cv2.imwrite(bar_save_path,barcode_img )  
    try:
        d = os.popen("D:\\zbar\\bin\\zbarimg.exe -q "+bar_save_path)#这里要根据自己的zbar调用路径做修改
        bar_info = d.readlines()[0]
        bar_info = bar_info.strip('\n').split(":")[1]  #读取命令行的输出
    except:
        bar_info = ""
       
    return bar_info 

 
def recogBar( img ):
    pts,bar = getCnt(img)
   
    if len(pts) == 0: 
        return  ""
  
    str_bar =""
    try:
        pst_order = orderPts(pts)   #对4个点顺序排好序
        res_img = four_point_transform( img , pst_order )  # 4点变换
        str_bar = recogBarcode(res_img)

    except:
        return  ""
 
    return str_bar

if  __name__=="__main__":
    img = cv2.imread("imgs/41.png")
    
    str_bar = recogBar( img )
    
    print( "识别结果：" , str_bar )

三、运行结果

代码将对一张倾斜了的条形码的输入图像，成功识别，将打印输出条码的内容，以及在代码根目录下生成矫正了的条码图像

输入图像

矫正了的图像

打印识别结果