1.欢迎点赞、关注、批评、指正，互三走起来，小手动起来！
2.了解、学习OCR相关技术知识领域，结合日常的场景进行测试、总结。如本文总结的html" title=flask>flask+paddlehtml" title=ocr>ocr+html" title=bootstrap>bootstrap搭建OCR文本推理WEB服务应用示例场景。
文章目录
- 1.代码结构
- 2.效果演示
- 3.接口返回
- 4.代码详情
- 4.1 `HTML`代码详情
- 4.2`Python`代码详情
- 5.PaddleOCR模型推理参数解释
- 6.后续展望
1.代码结构
- 如下图所示：
2.效果演示
- 详情如下：
3.接口返回
-
接口测试效果图
-
返回结果详情部分示例
{"error_code": "000000","error_msg": "识别成功","filename": "cstp2.png","recognize_time": "5890","result": [{"points": [[14.0,11.0],[108.0,11.0],[108.0,40.0],[14.0,40.0]],"score": 0.979973316192627,"text": "性别：男"},{"points": [[289.0,11.0],[400.0,9.0],[401.0,37.0],[290.0,39.0]],"score": 0.8993546962738037,"text": "住院号："},{"points": [[677.0,2.0],[713.0,12.0],[707.0,31.0],[672.0,21.0]],"score": 0.6370271444320679,"text": "贝别："}] }
4.代码详情
4.1 HTML代码详情
-
html"><!DOCTYPE html> <html> <meta charset="utf-8"> <head><title>OCR文字检测识别试运行系统</title><!--静态加载 样式--><link rel="stylesheet" href={{ url_for('static',filename='html" title=bootstrap>bootstrap-3.4.1/css/html" title=bootstrap>bootstrap.min.css') }}></link><link rel="stylesheet" href={{ url_for('static',filename='css/upload.css') }}></link><link rel="stylesheet" href={{ url_for('static',filename='css/36buttons.css') }}></link> </head> <body> <div class="header"><div class="title">【OCR】PP-OCRv3 文字检测识别试运行系统v0.3.0</div> </div><ul class="menu"><li><a href="/upload/">通用文本检测识别处理</a></li> </ul><div class="content"><!--上传图片文件--><div id="upload_file"><form id="fileForm" action="/upload/" method="POST" enctype="multipart/form-data"><div class="form-group"><input type="file" class="form-control" id="_upload_file" name="upload_file"> <!-- <label class="sr-only" for="upload_file">上传图片</label>--></br><button id="resetButton" name="resetButton" type="reset" class="button green">重置推理结果</button></div></form></div> </div> </div><div id="show" style="display: none;"><!--显示上传的图片--><div class="col-md-6" style="border: 2px solid #ddd;"></br><span class="label label-info" style="font-size: 24px;"><<<<<< 原始图片展示 >>>>>> </br></span><!--静态加载 图片, url_for() 动态生成路径 --></br><img id="src_pic_show" src="" alt="Image preview area..." title="preview-img" class="img-responsive"></div><div class="col-md-6" style="border: 2px solid #ddd;"><!--显示识别结果JSON报文列表--></br><span class="label label-info" style="font-size: 24px;"><<<<<< 推理结果详情 >>>>>> </br></span></br><!-- 结果显示区 --><div id="result_show" style="font-size: 28px;">客官，您提交的任务加急推理中......</div></div> </div> </body> </html> <script src="https://code.jquery.com/jquery-1.12.4.min.js"></script> <script src="http://malsup.github.io/jquery.form.js"></script><script type="text/javascript">var fileInput = document.querySelector('input[type=file]');var previewImg = document.querySelector('img');{#上传图片事件#}fileInput.addEventListener('change', function () {var file = this.files;var reader = new FileReader();//显示预览界面$("#show").css("display", "block");// 监听reader对象的的onload事件，当图片加载完成时，把base64编码賦值给预览图片reader.addEventListener("load", function () {previewImg.src = reader.result;}, false);// 调用reader.readAsDataURL()方法，把图片转成base64reader.readAsDataURL(file);//初始化输出结果信息$("#result_show").html("</br>客官，您提交的任务加急推理中......");{#上传图片识别表单事件，并显示识别结果信息#}{# ajaxSubmit 请求异步响应#}$("#fileForm").ajaxSubmit(function (data) {var inner = '<table border="1"> <thead> <tr> <th>序号</th> <th>文本目标</th> <th>置信度分数</th> </tr> </thead> <tbody>';//循环输出返回结果，响应识别结果为每行列表var inc = 1;for (var i in data['result']) {var text = data['result'][i]['text'];var score = data['result'][i]['score'];inner += "<tr><td>" + inc + "</td>" + "<td>" + text + "</td>" + "<td>" + score + "</td></tr>";inc += 1;}inner += '</tbody></table>'//清空输出结果信息$("#result_show").html("");//添加识别结果信息$("#show").append( inner );});}, false);document.getElementById('resetButton').addEventListener('click', function() {document.getElementById('src_pic_show').src = '';$('#result_show').empty();}); </script>
4.2Python代码详情
- 源代码
import json import os import time import numpy as np import pandas as pd from pyautogui import * from paddlehtml" title=ocr>ocr import PaddleOCR from PIL import Image, ImageDraw from html" title=flask>flask import Flask, render_template, request, jsonify # 应用名称，当前py名称，视图函数 app = Flask(__name__) # 相对路径 BASE_DIR = os.path.dirname(os.path.basename(__name__))# 上传文件路径 UPLOAD_DIR = os.path.join(os.path.join(BASE_DIR, 'static'), 'upload')def html" title=ocr>ocr_img2text( image ):result_dict = {'result': []}paddlehtml" title=ocr>ocr = PaddleOCR(det_model_dir='./inference/ch_PP-OCRv3_det_infer/',rec_model_dir='./inference/ch_PP-OCRv3_rec_infer/',cls_model_dir='./inference/ch_pphtml" title=ocr>ocr_mobile_v2.0_cls_infer/',use_angle_cls=True, lang="ch", use_gpu=True)if image == "":image = screenshot() # 使用pyautogui进行截图操作image = np.array(image)else:# 不为空就打开image = Image.open(image).convert('RGB')image = np.array(image) # 经提醒，需要添加arrayprint( image, type(image) )# 识别图片文件result0 = paddlehtml" title=ocr>ocr.html" title=ocr>ocr( image, cls=True )result = result0[0]# for line in result0:# for word in line:# print( word )for index in range(len(result)):line = result[index]p_dict = {}points = line[1]text = line[1][0]score = line[1][1]p_dict['points'] = pointsp_dict['text'] = textp_dict['score'] = scoreresult_dict['result'].append( p_dict )return result_dict@app.route('/') def upload_file():return render_template('upload.html')@app.route('/upload/', methods=['GET', 'POST']) def upload():if request.method == 'POST':# 每个上传的文件首先会保存在服务器上的临时位置，然后将其实际保存到它的最终位置。filedata = request.files['upload_file']upload_filename = filedata.filenameprint(upload_filename)# 保存文件到指定路径# 目标文件的名称可以是硬编码的，也可以从 request.files[file] 对象的 filename 属性中获取。# 但是，建议使用 secure_filename() 函数获取它的安全版本if not os.path.exists(UPLOAD_DIR):os.makedirs(UPLOAD_DIR)img_path = os.path.join(UPLOAD_DIR, upload_filename)filedata.save(img_path)start = time.time()# 打开图片img1 = Image.open(img_path)# 识别图片result_dict = html" title=ocr>ocr_img2text(img_path)# 识别时间end = time.time()recognize_time = int((end - start) * 1000)result_dict["filename"] = upload_filenameresult_dict["recognize_time"] = str(recognize_time)result_dict["error_code"] = "000000"result_dict["error_msg"] = "识别成功"return jsonify(result_dict)else:return render_template('upload.html')
5.PaddleOCR模型推理参数解释
- 参数详情
6.后续展望
- 持续改进优化该部分代码，并完善文档。欢迎交流。。。
- PaddleOCR模型推理参数解释
- https://www.cnblogs.com/xh2023/p/17642994.html
![[Qt][Qt窗口][上]详细讲解](https://i-blog.csdnimg.cn/direct/a7106275d43044bb817ecf0c9d23f412.png)