Python識別圖片中的文字

import os,glob
def photo_compression(original_imgage,tmp_image_path):
    '''圖片備份、壓縮；param original_imgage:原始圖片路徑；param tmp_imgage_path:臨時圖片路徑，備份路徑；return'''
    from PIL import Image
    img = Image.open(original_imgage)
    width,height = img.size
    while (width*height>4000000):#該數值壓縮后的圖片大約200多k
        width = width//2
        height = height//2
    e_img = img.resize((width,height),Image.BILINEAR)
    save_path = os.path.join(tmp_image_path,os.path.basename(original_imgage))
    e_img.save(save_path)
    return save_path

def ocr(original_image):
    '''使用百度OCR進行文字識別，支持JPG、JPEG、PNG、BMP格式；param original_image:待識別圖片；return'''
    from aip import AipOcr
    filename = os.path.basename(original_image)
    #輸入自己的百度ai賬號ID密碼：參考鏈接：https://m.toutiaocdn.com/i6704242394566492684/
    APP_ID = '******'
    API_KEY = '*******'
    SECRECT_KEY = '*********'

    client = AipOcr(APP_ID,API_KEY,SECRECT_KEY)

    with open(original_image,'rb') as picfile_read:
        img = picfile_read.read()
        print('正在識別圖片：{0}......'.format(filename))
        try:
            result = client.basicGeneral(img)#通用文字識別，50000次/天免費
        except:
            result = client.basicAccurate(img)#通用文字識別（高精度版），500次/天免費
    return result

def run_ocr(original_image,tmp_image_path,result_file_path='identify_results.txt'):
    '''主函數 批量執行圖片文本識別，結果存儲；original_image:原始圖片；tmp_image_path:臨時圖片；result_file_path:識別文字存儲文件；return'''

    if os.path.exists(result_file_path):#判斷是否存在歷史識別結果，若存在則刪除
        os.remove(result_file_path)
    if not os.path.exists(tmp_image_path):#判斷臨時圖片路徑是否存在，若不存在則創建
        os.mkdir(tmp_image_path)
    tmp_file_path = []#臨時文件路徑列表
    for picfile in glob.glob(original_image):#glob.glob的參數是一個只含有方括號、問號、正斜線的正則表達式
        tmp_file = photo_compression(picfile,tmp_image_path)
        tmp_file_path.append(tmp_file)
    for picfile in tmp_file_path:#遍歷所有文件，進行OCR識別 結果存儲
        result = ocr(picfile)
        lines = [text.get('words') + '\n' for text in result.get('words_result')]
        # lines = [text.get('words').encode('utf-8')+'\n' for text in result.get('words_result')]

        with open(result_file_path,'a+',encoding='utf-8') as fo:
            fo.writelines("="*100+'\n')
            fo.writelines("【識別圖片】：{0} \n".format(os.path.basename(picfile)))
            fo.writelines("【文本內容】： \n")
            fo.writelines(lines)
        os.remove(picfile)

if __name__ == '__main__':
    tmp_image_path = os.getcwd()+'\\tmp'
    original_image = os.getcwd() + '\\*.png'
    run_ocr(original_image,tmp_image_path)