1. 获取百度识字API

由于百度提供了免费的图片识字API，而且非常好用，所以直接使用百度的api来进行识字。

首先在百度AI开放平台注册一个账号，网址http://ai.baidu.com/，注册完之后进入右上角控制台，点击->"文字识别"

点击“创建应用”

创建一个识字的应用，接口选择“文字识别”，然后下面填个人，描述随便写一些

然后再应用主页中可以看到申请到的API key和Secret Key，该接口一天可以每天调用50000次，完全可以满足我们个人的需求

2. 调用百度API

拿到API_KEY和SECRET_KEY后获取token，然后通过token和接口来对图片进行识字，这里使用本地tmp.png作为目标图片进行识别。代码如下，想用的直接复制就好

baidu.py

"""

baidu.py

"""

import requests

import base64

API_KEY = '你的API_KEY'

SECRET_KEY = '你的SECRET_KEY'

def get_access_token():

# 获取token值

# client_id 为官网获取的API_KEY， client_secret 为官网获取的SECRET_KEY

host = 'https://aip.baidubce.com/oauth/2.0/token?grant_type=' + \

'client_credentials&client_id=' + API_KEY + \

'&client_secret=' + SECRET_KEY

response = requests.get(host)

access_token = response.json()['access_token']

return access_token

access_token = get_access_token()

def img2word():

# 将图片中的文字识别出来

global access_token

request_url = "https://aip.baidubce.com/rest/2.0/ocr/v1/general_basic"

# 二进制方式打开图片文件

f = open('./tmp.png', 'rb')

img = base64.b64encode(f.read())

params = {"image":img}

request_url = request_url + "?access_token=" + access_token

headers = {'content-type': 'application/x-www-form-urlencoded'}

response = requests.post(request_url, data=params, headers=headers)

print(response.json())

ret = [x['words'] for x in response.json()['words_result']]

return ret

3. 实现截图功能

识字部分已经完成，接下实现截图功能，我们想要像微信里面截图那样可以选中区域进行截图，代码如下

scrshoot.py

"""

scrshoot.py

"""

from PIL import ImageGrab

import tkinter

import ctypes

class CTkPrScrn:

def __init__(self):

self.__start_x, self.__start_y = 0, 0

self.__scale = 1

self.__win = tkinter.Tk()

self.__win.attributes("-alpha", 0.1) # 设置窗口半透明

self.__win.attributes("-fullscreen", True) # 设置全屏

self.__win.attributes("-topmost", True) # 设置窗口在最上层

self.__width, self.__height = self.__win.winfo_screenwidth(), self.__win.winfo_screenheight()

# 创建画布

self.__canvas = tkinter.Canvas(self.__win, width=self.__width, height=self.__height, bg="white")

self.__win.bind('', self.xFunc1) # 绑定鼠标左键点击事件

self.__win.bind('', self.xFunc1) # 绑定鼠标左键点击释放事件

self.__win.bind('', self.xFunc2) # 绑定鼠标左键点击移动事件

self.__win.bind('', lambda e: self.__win.destroy()) # 绑定Esc按键退出事件

user32 = ctypes.windll.user32

gdi32 = ctypes.windll.gdi32

dc = user32.GetDC(None)

widthScale = gdi32.GetDeviceCaps(dc, 8) # 分辨率缩放后的宽度

heightScale = gdi32.GetDeviceCaps(dc, 10) # 分辨率缩放后的高度

width = gdi32.GetDeviceCaps(dc, 118) # 原始分辨率的宽度

height = gdi32.GetDeviceCaps(dc, 117) # 原始分辨率的高度

self.__scale = width / widthScale

print(self.__width, self.__height, widthScale, heightScale, width, height, self.__scale)

self.__win.mainloop() # 窗口持久化

def xFunc1(self, event):

# print(f"鼠标左键点击了一次坐标是:x={g_scale * event.x}, y={g_scale * event.y}")

if event.state == 8: # 鼠标左键按下

self.__start_x, self.__start_y = event.x, event.y

elif event.state == 264: # 鼠标左键释放

if event.x == self.__start_x or event.y == self.__start_y:

return

im = ImageGrab.grab((self.__scale * self.__start_x, self.__scale * self.__start_y,

self.__scale * event.x, self.__scale * event.y))

imgName = 'tmp.png'

im.save(imgName)

print('保存成功')

self.__win.update()

self.__win.destroy()

self.__win.quit()

def xFunc2(self, event):

# print(f"鼠标左键点击了一次坐标是:x={self.__scale * event.x}, y={self.__scale * event.y}")

if event.x == self.__start_x or event.y == self.__start_y:

return

self.__canvas.delete("prscrn")

self.__canvas.create_rectangle(self.__start_x, self.__start_y, event.x, event.y, outline='red', tag="prscrn")

# 包装画布

self.__canvas.pack()

if __name__ == '__main__':

prScrn = CTkPrScrn()

4. tkiner实现简单UI

通过tk库简单实现一个窗口用于截图和显示文本，同时可以复制文本

front.py

"""

front.py

"""

import tkinter as tk

from baidu import img2word

from scrshoot import CTkPrScrn

def center_window(root, width, height):

# 将window设置在屏幕中间

screenwidth = root.winfo_screenwidth()

screenheight = root.winfo_screenheight()

size = '%dx%d+%d+%d' % (width, height, (screenwidth - width)/2, (screenheight - height)/2)

root.geometry(size)

def srcandserach():

global root, text_words

CTkPrScrn()

words = ' '.join(img2word())

print(words)

text_words.delete(0.0,tk.END)

text_words.insert(tk.INSERT,words)

root.update()

ft = ('Arial', '12') # 字体大小

root = tk.Tk()

root.title('识字搜索')

center_window(root, 800,300)

# 截图按钮

button_scr = tk.Button(root, text="截图", font=ft, command=srcandserach)

text_words = tk.Text(root, width=80,height=10, font=ft)

button_scr.grid(row=0, column=0)

text_words.grid(row=1, column=0)

root.mainloop()

5. 工具展示

识字ui

点击截图识字，可以识别出截图的文字信息，大功告成啦~

原文链接:https://blog.csdn.net/blowfire123/article/details/112079583