python 驗證碼識別示例（二） 復雜驗證碼識別
阿新 • • 發佈：2018-08-10
def code 分享 切割 edr convert pen https lis
在這篇博文中手把手教你如何去分割驗證，然後進行識別。
一：下載驗證碼
驗證碼分析，圖片上有折線，驗證碼有數字，有英文字母大小寫，分類的時候需要更多的樣本，驗證碼的字母是彩色的，圖片上有雪花等噪點，因此識別改驗證碼難度較大
二：二值化和降噪：
三： 切割：
四：分類：
五： 測試識別率
六：總結：
綜合識別率在70%左右，對於這個識別率我覺得還是挺高的，因為這個驗證碼的識別難度還是很大
代碼：
1. 下載圖片：
#-*-coding:utf-8-*-import requests def spider(): url = "https://www.epailive.com/basic/captcha?ran=0.22070346581876787" for i in range(1, 101): print("正在下載的張數是：",i) with open("./1__get_image/{}.png".format(i), "wb") as f: f.write(requests.get(url).content) spider()
二值化和降噪：
#-*-coding:utf-8-*- # coding:utf-8 import sys, os from PIL import Image, ImageDraw # 二值數組 t2val = {} def twoValue(image, G): for y in range(0, image.size[1]): for x in range(0, image.size[0]): g = image.getpixel((x, y)) if g > G: t2val[(x, y)]= 1 else: t2val[(x, y)] = 0 # 根據一個點A的RGB值，與周圍的8個點的RBG值比較，設定一個值N（0 <N <8），當A的RGB值與周圍8個點的RGB相等數小於N時，此點為噪點 # G: Integer 圖像二值化閥值 # N: Integer 降噪率 0 <N <8 # Z: Integer 降噪次數 # 輸出 # 0：降噪成功 # 1：降噪失敗 def clearNoise(image, N, Z): for i in range(0, Z): t2val[(0, 0)] = 1 t2val[(image.size[0] - 1, image.size[1] - 1)] = 1 for x in range(1, image.size[0] - 1): for y in range(1, image.size[1] - 1): nearDots = 0 L = t2val[(x, y)] if L == t2val[(x - 1, y - 1)]: nearDots += 1 if L == t2val[(x - 1, y)]: nearDots += 1 if L == t2val[(x - 1, y + 1)]: nearDots += 1 if L == t2val[(x, y - 1)]: nearDots += 1 if L == t2val[(x, y + 1)]: nearDots += 1 if L == t2val[(x + 1, y - 1)]: nearDots += 1 if L == t2val[(x + 1, y)]: nearDots += 1 if L == t2val[(x + 1, y + 1)]: nearDots += 1 if nearDots < N: t2val[(x, y)] = 1 def saveImage(filename, size): image = Image.new("1", size) draw = ImageDraw.Draw(image) for x in range(0, size[0]): for y in range(0, size[1]): draw.point((x, y), t2val[(x, y)]) image.save(filename) for i in range(1, 101): path = "1__get_image/" + str(i) + ".png" image = Image.open(path) image = image.convert(‘L‘) twoValue(image, 198) clearNoise(image, 3, 1) path1 = "2__erzhihua_jiangzao/" + str(i) + ".jpg" saveImage(path1, image.size)
三：切割驗證碼：
#-*-coding:utf-8-*- from PIL import Image def smartSliceImg(img, outDir, ii,count=4, p_w=3): ‘‘‘ :param img: :param outDir: :param count: 圖片中有多少個圖片 :param p_w: 對切割地方多少像素內進行判斷 :return: ‘‘‘ w, h = img.size pixdata = img.load() eachWidth = int(w / count) beforeX = 0 for i in range(count): allBCount = [] nextXOri = (i + 1) * eachWidth for x in range(nextXOri - p_w, nextXOri + p_w): if x >= w: x = w - 1 if x < 0: x = 0 b_count = 0 for y in range(h): if pixdata[x, y] == 0: b_count += 1 allBCount.append({‘x_pos‘: x, ‘count‘: b_count}) sort = sorted(allBCount, key=lambda e: e.get(‘count‘)) nextX = sort[0][‘x_pos‘] box = (beforeX, 0, nextX, h) img.crop(box).save(outDir + str(ii) + "_" + str(i) + ".png") beforeX = nextX for ii in range(1, 101): path = "2__erzhihua_jiangzao/" + str(ii) + ".jpg" img = Image.open(path) outDir = ‘3__qiege/‘ smartSliceImg(img, outDir, ii,count=4, p_w=3)
四：訓練：
#-*-coding:utf-8-*- import numpy as np import os import time from PIL import Image from sklearn.externals import joblib from sklearn.neighbors import KNeighborsClassifier def load_dataset(): X = [] y = [] for i in "23456789ABVDEFGHKMNPRSTUVWXYZ": target_path = "fenlei/" + i print(target_path) for title in os.listdir(target_path): pix = np.asarray(Image.open(os.path.join(target_path, title)).convert(‘L‘)) X.append(pix.reshape(25 * 30)) y.append(target_path.split(‘/‘)[-1]) X = np.asarray(X) y = np.asarray(y) return X, y def check_everyone(model): pre_list = [] y_list = [] for i in "23456789ABCDEFGHKMNPRSTUVWXYZ": part_path = "part/" + i for title in os.listdir(part_path): pix = np.asarray(Image.open(os.path.join(part_path, title)).convert(‘L‘)) pix = pix.reshape(25 * 30) pre_list.append(pix) y_list.append(part_path.split(‘/‘)[-1]) pre_list = np.asarray(pre_list) y_list = np.asarray(y_list) result_list = model.predict(pre_list) acc = 0 for i in result_list == y_list: print(result_list,y_list,) if i == np.bool(True): acc += 1 print(acc, acc / len(result_list)) X, y = load_dataset() knn = KNeighborsClassifier() knn.fit(X, y) joblib.dump(knn, ‘yipai.model‘) check_everyone(knn)
六：測試：
# -*- coding: utf-8 -*- import numpy as np from PIL import Image from sklearn.externals import joblib import os target_path = "1__get_image/" source_result = [] for title in os.listdir(target_path): source_result.append(title.replace(‘.png‘,‘‘)) def predict(model): predict_result = [] for q in range(1,101): pre_list = [] y_list = [] for i in range(0,4): part_path = "part1/" + str(q) + "_" + str(i) + ".png" # print(part_path) pix = np.asarray(Image.open(os.path.join(part_path))) pix = pix.reshape(25 * 30) pre_list.append(pix) y_list.append(part_path.split(‘/‘)[-1]) pre_list = np.asarray(pre_list) y_list = np.asarray(y_list) result_list = model.predict(pre_list) print(result_list,q) predict_result.append(str(result_list[0] + result_list[1] + result_list[2] + result_list[3])) return predict_result model = joblib.load(‘yipai.model‘) predict_result = predict(model) # print(source_result) # print(predict_result)
python 驗證碼識別示例（二） 復雜驗證碼識別