Python 新手实战之机器学习实现简单验证码识别(二)：图片二值化除去噪点

图片二值化去噪点

import os
from PIL import Image
import numpy as np
import pandas as pd
from pandas import DataFrame
from threading import Thread
import queue
import time


class Clean(Thread):
    def __init__(self, que, out):
        """
        :param que:
        :type data: DataFrame
        """
        super(Clean, self).__init__()
        self.que = que
        self.out = out
        self.start()

    def to_grey(self, image):
        self.width, self.height = image.size
        data = np.asarray(image).reshape((self.width * self.height, 3))
        new_image = []
        for point in data:
            gp = (15 * point[0] + 75 * point[1] + 38 * point[2]) >> 7        #计算灰度值
            new_image.append(0 if gp < 180 else 255)                      #二值化处理（有更好的算法，这里直接用了180）
        return np.asarray(new_image).reshape((self.height, self.width))  #图片还原返回

    def clean_point(self, grey0, degree=200):
        grey = grey0.copy()                                   #创建一个副本 以免去除噪点后影响后续点的计算
        for x in range(1, len(grey) - 1):
            for y in range(1, len(grey[x]) - 1):
                mar = grey[x - 1:x + 2, y - 1: y + 2]        #取得点（x,y）及周围八个点
                if mar[mar > degree].size > 5:              #灰度值大于200的点的个数大于5就认为是噪点
                    grey0[x][y] = 255
        return grey0[1:-1, 1:-1]                           #返回原图 除去边框，上面没有除掉边框的噪点，并且边框用不到，直接舍去


    def clean(self, filename, filepath):
        image = Image.open(filepath)
        grey = self.to_grey(image)
        smooth = self.clean_point(grey)


    def run(self):
        while True:
            try:
                filename, filepath = self.que.get(block=False)
                res = self.clean(filename, filepath)
                if res == 0:
                    break
                self.que.task_done()
            except Exception as e:
                print(e)
                break


class Manager:
    def __init__(self, thread=1):
        self.threads = []
        files = os.listdir('./image')      #找出图片文件名
        filepaths = [os.path.join('./image/', file) for file in files] #构建图片路径
        filenames = [file[6:-4] for file in files]   #从文件名中获得验证码原文
        self.que = queue.Queue()               #创建任务队列和输出队列
        self.out = queue.Queue()
        self.init_work_queue(filenames, filepaths)  #初始化任务队列
        self.init_thread(thread)                     #初始化线程
        self.wait_allcomplete()

    def init_thread(self, thread):
        for i in range(thread):
            self.threads.append(Clean(self.que, self.out))

    def init_work_queue(self, filenames, filepaths):
        for i in range(len(filenames)):
            self.que.put((filenames[i], filepaths[i]))

    def wait_allcomplete(self):
        for item in self.threads:
            if item.isAlive():
                item.join()


if __name__ == '__main__':
    manager = Manager(2)    #初始化线程管理  2个线程
    data = pd.DataFrame()
    while True:
        try:
            data = data.append(manager.out.get(block=False), ignore_index=True)
        except:
            break
    data.to_csv('./csv/res.csv')

后续图片分割 数据处理在下一节完善