HyperLPR/upload.py

#coding=utf-8
from flask import Flask, render_template, request
#导入flaks 框架 渲染模板 和请求框架
from werkzeug import secure_filename

import cv2
import numpy as np

#导入opencv

from hyperlpr import pipline
#导入车牌识别库


app = Flask(__name__)
#设置App name


def recognize(filename):
    image = cv2.imread(filename)
    #通过文件名读入一张图片 放到 image中
    return pipline.RecognizePlateJson(image)
    #识别一张图片并返回json结果

#识别函数

import base64


def recognizeBase64(base64_code):
    file_bytes = np.asarray(bytearray(base64.b64decode(base64_code)),dtype=np.uint8)
    image_data_ndarray = cv2.imdecode(file_bytes,1)
    return pipline.RecognizePlateJson(image_data_ndarray)


import time

@app.route('/uploader', methods=['GET', 'POST'])#设置请求路由
def upload_file():
    if request.method == 'POST':
        #如果请求方法是POST
        f = request.files['file']
        f.save("./images_rec/"+secure_filename(f.filename))
        #保存请求上来的文件
        t0 = time.time()
        res = recognize("./images_rec/"+secure_filename(f.filename))
        print "识别时间",time.time() - t0
        return res
        #返回识别结果

        # return 'file uploaded successfully'
    return render_template('upload.html')



if __name__ == '__main__':
    #入口函数

    app.run("0.0.0.0",port=8000)
    #运行app 指定IP 指定端口