却是如前所述疗养院传染科的需求，中后期会有大批的样品相片，需要依次分类去辨识校正，透过程序手动进行分类增加下育苗进行分类的成本，简单标识符同时实现下。
路子
先透过百度普通OCR辨识相片文本信息，对文本信息结点，看与否有关连的关键字信息。如果辨识不行，在邻近地区的tesseract在辨识一场，辨识不出在标准化放两个产品目录中，育苗处理。
进行分类前：
进行分类后：
标识符同时实现
#!/usr/bin/env python # -*- coding: utf-8 -*- # @Time : 2022-7-8 下午 9:09 # @Author : damon.guo # @File : picToClass.py # @Software: PyCharm import cv2 from PIL import Image import pytesseract
importos,sys import shutil from aip import AipOcr gettaglist = ["诊断书", "调查报告", "疗养院", "卫生所", "服务中心"] # 辨识关键字 def picToClass(picturePath): # imageCode = cv2.imread(r"F:\xbl\11\\12.jpg") # Nanded，二值化 # picturePath=r"F:\xbl\11\\preclass\\1.jpg" # Nanded，二值化 image = cv2.imread(picturePath) gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY) gray = cv2.GaussianBlur(gray, (5, 5), 0) edged = cv2.Canny(gray, 75, 200) tmpfilename = r"F:\xbl\11\\class\\tmp.jpg" # 当文件 cv2.imwrite(tmpfilename, edged) text = pytesseract.image_to_string(Image.open(tmpfilename), lang=chi_sim) print("二次辨识",picturePath) for i in text.split("\n"): index = getTagIndex(i) if index != -1: news = i[:index] print("nes", news) return news.strip(" ").strip(":").strip("%") # 却是辨识不出来，标准化放两个产品目录 return "other" def getclient(APP_ID, API_KEY, SECRET_KEY): """ 你的 APPID AK SK """ # APP_ID = 你的 App ID # API_KEY = 你的 Api Key # SECRET_KEY = 你的 Secret Key client = AipOcr(APP_ID, API_KEY, SECRET_KEY) return client def get_file_content(imagepath): with open(imagepath, rb) as fp: return fp.read() def ocrsdk(imagepath): #先采用百度sdk辨识 client = getclient("xx", "xx", "xx") image = get_file_content(imagepath) # 必填参数 options = {} # options["templateSign"] = "" # 模板id # options["detect_direction"] = "true" #与否手动转向 # options["probability"] = "true" # options["language_type"] = "CHN_ENG" # 辨识语言 res_image = client.basicAccurate(image,options) print("辨识相片：",image) tag = baiduocrreslut(res_image) if tag == "other": # 无法辨识在 # 在邻近地区ocr辨识一场 tag = picToClass(imagepath) return tag def getTagIndex(strs): # 辨识关键字在位置，取到索引，并向后多取2位 for i in gettaglist: if i in strs: index = strs.index(i) return index+2 return -1 def
copyFile(src,dst): shutil.copy(src, dst) def baiduocrreslut(res): # 百度辨识结果解析，
wordlist = res["words_result"] for w in wordlist: for i in gettaglist: if i in w["words"]: print(w["words"]) if len(w["words"]) >= 4: return w[words] return "other" def main(): path = r"F:\xbl\preclass" # 待进行分类相片产品目录 classpath = r"F:\xbl\class" # 进行分类后主产品目录 listf = os.listdir(path) for i in listf: imagepath = os.path.join(path,i) tag = ocrsdk(imagepath) tagpath = os.path.join(classpath,tag) if not os.path.
exists(tagpath): os.makedirs(tagpath) newimagepath = os.path.join(tagpath,i) if not os.path.exists(newimagepath): shutil.copy(imagepath,newimagepath) if __name__ == "__main__": main()
