opencv实战-识别信用卡
一、识别信用卡步骤
1、读取模板文件
2、对模板文件进行灰度处理
3、对模板文件进行二值处理
4、对模板进行轮廓检测并计算外接矩形
5、读取行用卡图片
6、信用卡灰度处理
7、对信用卡二值处理
8、对信用卡进行梯度计算
9、对信用卡进行轮廓检测
10、对信用卡检测出的轮廓进行二值处理
11、对每个数字进行切分
12、模板匹配
二、参考代码
import numpy as np import argparse import cv2 # 定义图片显示 def cv_show(name, img): cv2.imshow(name, img) cv2.waitKey(0) cv2.destroyAllWindows() def takeSecond(elem): return elem[0] # 定义模板的排序,默认从左到右 def sort_contours(cnts, method="left-to-right"): reverse = False if method == "right-to-left" or method == "bottom-to-top": reverse = True if method == "top-to-bottom" or method == "bottom-to-top": i = 1 # 算出外接矩形的的坐标x,y,w,h boundingBoxes = [cv2.boundingRect(c) for c in cnts] # 将cnts和外接矩形的坐标轴根据第boundingBoxes[1][i]进行排序 (cnts, boundingBoxes) = zip(*sorted(zip(cnts, boundingBoxes), key=lambda b: b[1][i], reverse=reverse)) return cnts, boundingBoxes # 对图像进行变换 def resize(image, width=None, height=None, inter=cv2.INTER_AREA): dim = None # 将图像二维数组分开为h、w (h, w) = image.shape[:2] # 如果width、height不带参数则直接返回图片 if width is None and height is None: return image # 如果w为空,则使用((height/h)*w,height) if width is None: r = height / float(h) dim = (int(w * r), height) else: # 如果height为空,则使用(width,(width/w)*h) r = width / float(w) dim = (width, int(h * r)) resized = cv2.resize(image, dim, interpolation=inter) return resized # 读取模板文件 img = cv2.imread('ocr_a_reference.png') # 将模板文件置为灰度图 ref = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) # 阈值处理,二值化 ref = cv2.threshold(ref, 10, 255, cv2.THRESH_BINARY_INV)[1] # 轮廓检测,只检测最外面的轮廓,压缩水平的、垂直的和斜的部分,也就是说函数保留他们的重点部分 contours, hierarchy = cv2.findContours(ref.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE) # 在原图像画出检测出的轮廓 cv2.drawContours(img, contours, -1, (0, 0, 255), 2) # 排序,从左到右,从上到下 # refCnts = sort_contours(contours, method="left-to-right")[0]# 算出外接矩形的的坐标x,y,w,h boundingBoxes = [cv2.boundingRect(c) for c in contours] # 将cnts和外接矩形的坐标轴进行排序 refCnts = sorted(boundingBoxes, key=takeSecond, reverse=False) digits = {} # 遍历每一个模板轮廓 for (i, c) in enumerate(refCnts): (x, y, w, h) = c roi = ref[y:y + h, x:x + w] roi = cv2.resize(roi, (60, 60)) # 每一个数字对应每一个模板 digits[i] = roi cv_show('', roi) # 初始化卷积核 rectKernel = cv2.getStructuringElement(cv2.MORPH_RECT, (9, 3)) sqKernel = cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5)) # 读取输入行用卡图像,预处理 image = cv2.imread('credit_card_01.png') image = resize(image, width=300) # 将信用卡置为灰度图 gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY) # 礼帽操作,突出更明亮的区域 tophat = cv2.morphologyEx(gray, cv2.MORPH_TOPHAT, rectKernel) # 进行梯度计算 gradX = cv2.Sobel(tophat, ddepth=cv2.CV_32F, dx=1, dy=0, ksize=-1) sobelx = cv2.convertScaleAbs(gradX) grady = cv2.Sobel(tophat, ddepth=cv2.CV_32F, dx=0, dy=1, ksize=-1) sobely = cv2.convertScaleAbs(grady) sobelxy = cv2.addWeighted(sobelx, 0.5, sobely, 0.5, 0) # 通过闭操作(先膨胀,再腐蚀)将数字连在一起 sobelxy = cv2.morphologyEx(sobelxy, cv2.MORPH_CLOSE, rectKernel) # 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 sobelxy = cv2.threshold(sobelxy, 0, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1] # 通过闭操作(先膨胀,再腐蚀) thresh = cv2.morphologyEx(sobelxy, cv2.MORPH_CLOSE, sqKernel) # 计算轮廓 threshCnts, hierarchy = cv2.findContours(thresh.copy(), cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) cur_img = image.copy() # 把所有轮廓画在原图上 cv2.drawContours(cur_img, threshCnts, -1, (0, 0, 255), 3) locs = [] # 遍历轮廓 for (i, c) in enumerate(threshCnts): # 计算外接矩形 (x, y, w, h) = cv2.boundingRect(c) # 计算长宽比 ar = w / float(h) # 选择合适的区域,这里的基本都是四个数字一组 if ar > 2 and ar < 4: if (w > 38 and w < 62) and (h > 10 and h < 24): # 符合的留下来 locs.append((x, y, w, h)) # 将符合的轮廓从左到右排序 locs = sorted(locs, key=lambda x: x[0]) output = [] # 遍历每一个轮廓中的数字 for (i, (gX, gY, gW, gH)) in enumerate(locs): groupOutput = [] # 根据坐标在原图里面提取每一个组 group = gray[gY - 5:gY + gH + 5, gX - 5:gX + gW + 5] group = cv2.resize(group, (60, 60)) group = group[0:56, 0:60] # 阈值处理 group = cv2.threshold(group, 0, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1] # 计算每一组的轮廓 digitCnts, hierarchy = cv2.findContours(group.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE) # 将轮廓从左到右排序 digitCnts = [cv2.boundingRect(c) for c in digitCnts] # 将cnts和外接矩形的坐标轴根据第boundingBoxes[1][i]进行排序 digitCnts = sorted(digitCnts, key=takeSecond, reverse=False) # 计算每一组中的每一个数值 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) = c roi = group[y:y + h, x:x + w] roi = cv2.resize(roi, (60, 60)) # cv_show('roi', roi) # 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))) # 画出来0 cv2.rectangle(image, (gX - 5, gY - 5), (gX + gW + 5, gY + gH + 5), (0, 0, 255), 1) cv2.putText(image, "".join(groupOutput), (gX, gY - 15), cv2.FONT_HERSHEY_SIMPLEX, 0.65, (0, 0, 255), 2) cv_show('',image) # 得到结果 output.extend(groupOutput) print("Credit Card #: {}".format("".join(output))) cv2.imshow("Image", image) cv2.waitKey(0)