python EasyOCR庫實例用法介紹
說明
1、EasyOCR是一個用python編寫的OCR三方庫。可以在python中調用,用來識別圖像中的文字,並輸出為文本。
2、支持80多種語言的識別,識別精度高,甚至要超過PaddleOCR。
安裝命令
pip install easyocr
代碼實現
import easyocr #設置識別中英文兩種語言 reader = easyocr.Reader(['ch_sim','en'], gpu = False) # need to run only once to load model into memory result = reader.readtext(r"d:\Desktop\4A34A16F-6B12-4ffc-88C6-FC86E4DF6912.png", detail = 0) print(result)
實例擴展:
圖文提取的代碼
from pathlib import Path import easyocr file_url = r'識別圖片.jpg' # 需識別的圖片 split_symbol = ' ' # 默認空格為分隔符 row_space = 15 # 默認字符高度為15px,當識別出來的字符間距超過這個數值時會換行。 def make_reader(): # 將模型加載到內存中。模型文件地址 C:\Users\用戶\.EasyOCR\model reader = easyocr.Reader(['ch_sim', 'en']) return reader def change_to_character(file_url, reader, split_symbol=' ', row_space=15, save_dir='.'): with open(file_url, "rb") as img: img_b = img.read() result = reader.readtext(img_b) result.sort(key=lambda x: x[0][0][1]) # 按豎直方向,進行排序==>進行分行處理。 # for i in result: # print(i) # print('='*100) # 按行進行分組 content = [] item = [result[0]] # 首先放入第一個元素 for i in result[1:]: if row_space >= i[0][0][1] - item[-1][0][0][1] >= 0: item.append(i) else: content.append(item) item = [i] content.append(item) filemane = Path(file_url).name.split('.')[0] with open(f'{save_dir}/{filemane}.txt', "w", encoding='utf8') as t: for i in content: # i 為每一行的內容 i.sort(key=lambda x: x[0][0][0]) # 對每行的內容進行先後排序 for r in i: # print(r) t.write(r[1] + split_symbol) t.write("\n") return content if __name__ == "__main__": change_to_character(file_url, make_reader())
UI 界面的代碼
import tkinter as tk from tkinter import filedialog from PIL import Image, ImageTk from pathlib import Path from character import change_to_character, make_reader from threading import Thread import time # class Showing(tk.Frame): # def __init__(self, master=None): # super().__init__(master) # self.master = master # self.pack() # # self.img = tk.PhotoImage(file=r"C:\Users\yanhy\Desktop\捕獲22.PNG") # self.create_widgets() # # def create_widgets(self): # self.img = tk.PhotoImage(file=r"C:\Users\yanhy\Desktop\捕獲22.PNG") # self.img_wig = tk.Label(self, image=self.img) # self.img_wig.pack() # 最外層窗口設置 root = tk.Tk() root.title('圖片文字識別程序 聯系:[email protected]') window_x = root.winfo_screenwidth() window_y = root.winfo_screenheight() WIDTH = 1200 HEIGHT = 750 x = (window_x - WIDTH) / 2 # 水平居中 y = (window_y - HEIGHT) / 3 # 垂直偏上 root.geometry(f'{WIDTH}x{HEIGHT}+{int(x)}+{int(y)}') root.resizable(width=False, height=False) # 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》 Row_space = 15 File_url_list = [] Img_type = ['.jpg', '.jpeg', '.png', '.gif'] Split_symbol = ' ' # 間隔符。 Save_dir = Path.cwd().joinpath('img_to_word') if Save_dir.is_dir(): pass else: Path.mkdir(Save_dir) # 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》 def test(): print(f'{Row_space=}') def choose_file(): # 獲取導入的圖片路徑地址 global show_img, img_label, text, File_url_list filenames = filedialog.askopenfilenames() if len(filenames) == 1 and len(File_url_list) == 0: # 單張圖片導入,顯示圖片 if Path(filenames[0]).suffix.lower() in Img_type: # 判斷是否圖片類型 File_url_list = list(filenames) try: if text.winfo_exists(): text.destroy() except NameError as e: print(f'choose_file提示:張圖片導入錯誤>>> {e}') try: if img_label.winfo_exists(): img_label.destroy() except NameError as e: print(f'choose_file提示:單張圖片導入錯誤>>> {e}') img = Image.open(File_url_list[0]).resize((560, 660)) # print(img.size) show_img = ImageTk.PhotoImage(image=img) img_label = tk.Label(f_left, image=show_img) img_label.pack() else: print('導入的是非圖像格式') else: # 多張圖片導入,顯示列表。 try: if img_label.winfo_exists(): img_label.destroy() except NameError as e: print(f'提示:多張圖片導入錯誤>>> {e}') try: if text.winfo_exists(): text.destroy() except NameError as e: print(f'提示:多張圖片導入錯誤>>> {e}') text = tk.Text(f_left, spacing1=5, spacing3=5) text.pack(fill='both', expand=True) for i in filenames: if Path(i).suffix.lower() in Img_type: File_url_list.append(i) else: pass File_url_list = set(File_url_list) for i in list(File_url_list): # 把文件寫入到文本框中 text.insert('end', str(list(File_url_list).index(i)+1) + ": " + i + "\n") File_url_list = list(File_url_list) print(f'{File_url_list=}') def choose_dir(): global show_img, img_label, text, File_url_list directoryname = filedialog.askdirectory() print(f'{directoryname=}') try: if img_label.winfo_exists(): img_label.destroy() except NameError as e: print(f'choose_dir提示:多張圖片導入錯誤>>> {e}') try: if text.winfo_exists(): text.destroy() except NameError as e: print(f'choose_dir提示:多張圖片導入錯誤>>> {e}') text = tk.Text(f_left, spacing1=5, spacing3=5) text.pack(fill='both', expand=True) for i in Path(directoryname).iterdir(): # 獲取文件夾下的所有文件。 if Path(i).suffix.lower() in Img_type: File_url_list.append(i.as_posix()) # as_posix() 把Path型轉為字符串。 else: pass File_url_list = set(File_url_list) for i in list(File_url_list): # 把文件寫入到文本框中 text.insert('end', str(list(File_url_list).index(i) + 1) + ": " + i + "\n") File_url_list = list(File_url_list) print(f'{File_url_list=}') def clear_file_list(): global File_url_list File_url_list.clear() try: if img_label.winfo_exists(): img_label.destroy() except NameError as e: print(f'clear_file_list提示:清空錯誤>>> {e}') try: if text.winfo_exists(): text.destroy() except NameError as e: print(f'clear_file_list提示:清空錯誤錯誤>>> {e}') def get_entry1(): # 設置換行間距變量值 global Row_space num = entry1.get() if num.isdigit(): if int(num) > 0: Row_space = int(num) else: entry1.delete(0, "end") entry1.insert(0, 15) Row_space = 15 def set_split_symbol(): global Split_symbol Split_symbol = entry2.get() print(f'{Split_symbol=}') def do_change(): if File_url_list: v.set("文字提取中,請稍後……") button_do.config(state='disable') # 使按鈕不可用。 # ======================================== def main(): reader = make_reader() for i in File_url_list: content = change_to_character(i, reader, row_space=Row_space, split_symbol=Split_symbol, save_dir=Save_dir) read_text.delete(1.0, "end") for c in content: # i 為每一行的內容 c.sort(key=lambda x: x[0][0][0]) # 對每行的內容進行先後排序 for r in c: # print(r) read_text.insert('end', r[1] + Split_symbol) read_text.insert('end', "\n") v.set("文字提取結束。") button_do.config(state='normal') # 恢復按鈕可用。 # ======================================== t = Thread(target=main, daemon=True) t.start() else: v.set("請先選擇圖片!") def join_file(): v.set("文件開始合並。") filst = list(Path(Save_dir).iterdir()) # 獲取文件夾中所有的文本文件。 with open(f'{Save_dir}/合並文件.txt', 'w', encoding='utf8') as join_f: for f in filst: with open(f, 'r', encoding='utf8') as r_f: read_con = r_f.read() join_f.write(f.name+'\n'+read_con + '\n\n') time.sleep(1) v.set("文件合並完畢。") # 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》 f_top = tk.Frame(root, height=65, width=1100, bd=1, relief="flat") # "sunken" "raised","groove" 或 "ridge" f_top.pack_propagate(False) # 如果不加這個參數,當Frame框架中加入部件時,會自動變成底層窗口,自身的特性會消失。 f_top.pack(side='top', pady=5) f_left = tk.Frame(root, height=660, width=560, bd=1, relief="groove") f_left.pack_propagate(False) f_left.pack(side='left', padx=20) f_right = tk.Frame(root, height=660, width=560, bd=1, relief="groove") f_right.pack_propagate(False) f_right.pack(side='left', padx=20) read_text = tk.Text(f_right, spacing1=5, spacing3=5) read_text.pack(fill='both', expand=True) # 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》 button_choose_file = tk.Button(f_top, text='選擇圖片', command=choose_file) button_choose_file.pack(side='left', padx=10, ipadx=5) button_choose_file = tk.Button(f_top, text='選擇文件夾', command=choose_dir) button_choose_file.pack(side='left', padx=10, ipadx=5) button_clear_file = tk.Button(f_top, text='清空選擇', bg='#FFEF2F', command=clear_file_list) button_clear_file.pack(side='left', padx=5, ipadx=5) # 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》 f_row_content = tk.Frame(f_top, height=50, width=300, bg="#D1D4D0", relief="flat") # "sunken" "raised","groove" 或 "ridge" f_row_content.pack_propagate(False) f_row_content.pack(side='left', padx=15) button_set_row_height = tk.Button(f_row_content, text='設置行間距', command=get_entry1) button_set_row_height.pack(side='left', ipadx=3, padx=3) entry1 = tk.Entry(f_row_content, font=('', 18), width=3) entry1.insert(0, 15) entry1.pack(padx=5, side='left') tk.Label(f_row_content, justify='left', text='填入像素值,設置換行間距。\n默認15個像素。').pack(side='left') # 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》 f_split = tk.Frame(f_top, height=50, width=215, bg="#D1D4D0", relief="flat") # "sunken" "raised","groove" 或 "ridge" f_split.pack_propagate(False) f_split.pack(side='left', padx=4) button_split = tk.Button(f_split, text='設置分隔符', command=set_split_symbol) button_split.pack(side='left', ipadx=3, padx=3) entry2 = tk.Entry(f_split, font=('', 18), width=3) entry2.insert(0, ' ') entry2.pack(padx=5, side='left') tk.Label(f_split, justify='left', text='默認一個空格').pack(side='left') # 《《《《《《《《《《《《《《《《《《《《《《 提取 合並文件 》》》》》》》》》》》》》》》》》》》》》》》》》 button_do = tk.Button(f_top, text='開始提取', bg='#4AB0FF', command=do_change) button_do.pack(side='left', padx=10, ipadx=2) button_join = tk.Button(f_top, text='合並文件', command=join_file) button_join.pack(side='left', padx=5, ipadx=2) v = tk.StringVar() v.set('info……') tk.Label(f_top, bg='#2EBD1D', justify='left', textvariable=v).pack(side='left') # 《《《《《《《《《《《《《《《《《《《《《《 右鍵菜單 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》 def copy_text(): read_text.event_generate("<<Copy>>") menubar = tk.Menu(tearoff=False) # root['menu'] = menubar # 沒有把這個 菜單部件 加入到 root 窗口的菜單屬性中,所以它不會在root窗口的頂部顯示。 menubar.add_command(label='復制', command=copy_text) def show_menu(event): """用 菜單部件 的 post 方法展示菜單""" menubar.post(event.x_root, event.y_root) read_text.bind('<Button-3>', show_menu) # 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》 root.mainloop()
到此這篇關於python EasyOCR庫實例用法介紹的文章就介紹到這瞭,更多相關python EasyOCR庫是什麼內容請搜索WalkonNet以前的文章或繼續瀏覽下面的相關文章希望大傢以後多多支持WalkonNet!
推薦閱讀:
- Python GUI之如何使用tkinter控件
- python tkinter庫的Text記錄點擊路經和刪除記錄詳情
- Python GUI 圖形用戶界面
- Python基礎之tkinter圖形化界面學習
- Python GUI編程詳解