devbox
凡人多烦事01
这个屌丝很懒，什么也没留下！
关注作者
热门标签
热门文章
当前位置:   article > 正文

【ZED2-3】python同时实现websockets与flask_socketserver和flask怎么同时运行

作者：凡人多烦事01 | 2024-03-21 09:37:53

赞

踩

socketserver和flask怎么同时运行

上篇博客【ZED-2】基于WebSockets库，通过多线程方式推送数据流_WXG1011的博客-CSDN博客采用多线程方式，基于websockets库实现数据流的推送，这篇博客主要实现将前端（nx板）代码移植到flask框架中，通过http方式展示，便于其他客户端访问，主要难点在于flask app的ip地址与websockets ip地址冲突。

基础介绍-flask框架

注：该Demo解决了通过http访问时，视频流卡顿的问题。

flask_demo.py

  1. # 测试rtsp视频流在前端的稳定性
  2. from flask import Flask, render_template, Response
  3. import cv2
  4. import queue
  5. import threading
  6. import time
  7.  
  8. app = Flask(__name__)
  9.  
  10.  
  11. @app.route('/video_feed')
  12. def video_feed():
  13.     # 通过将一帧帧的图像返回，就达到了看视频的目的。multipart/x-mixed-replace是单次的http请求-响应模式，如果网络中断，会导致视频流异常终止，必须重新连接才能恢复
  14.     return Response(Display(), mimetype='multipart/x-mixed-replace; boundary=frame')
  15.  
  16. @app.route('/')
  17. def index():
  18.     return render_template('index.html')
  19.  
  20. q = queue.Queue()
  21.  
  22. def Receive():
  23.     print("start Reveive")
  24.     cap = cv2.VideoCapture(0)
  25.     # cap = cv2.VideoCapture("rtsp://admin:sdiit888@10.1.93.5/Streaming/Channels/1")
  26.     while True:
  27.         q.put(cap.read()[1])
  28.         # 移除队列中的旧图
  29.         q.get() if q.qsize() > 1 else time.sleep(0.01)
  30.  
  31. def Display():
  32.     print("Start Displaying")
  33.     while True:
  34.         if q.empty() != True:
  35.             frame = q.get()
  36.             ret, buffer = cv2.imencode('.jpg', frame)
  37.             frame = buffer.tobytes()
  38.             # 使用yield语句，将帧数据作为响应体返回，content-type为image/jpeg
  39.             yield (b'--frame\r\n'
  40.                    b'Content-Type: image/jpeg\r\n\r\n' + frame + b'\r\n')
  41.  
  42. if __name__ == '__main__':
  43.     threads = [threading.Thread(target=Receive), threading.Thread(target=Display)]
  44.     for t in threads:
  45.         t.start()
  46.     for t in threads:
  47.         t.join()
  48.     threading.Thread(target=app.run(debug=True)).start()

index.html

  1. <!DOCTYPE html>
  2. <html lang="en">
  3. <head>
  4.     <meta charset="UTF-8">
  5.     <title>AI功能展示</title>
  6.     <style>
  7.         .divcss5{text-align: center}
  8.     </style>
  9. </head>
  10. <body>
  11.     <h1 align="center"><font size="6" face="arial" color="black">视频展示</font></h1>
  12.     <div class="divcss5">
  13.         <img src="{{ url_for('video_feed') }}" height="900px" width="3400px">#}
  14.     </div>
  15.     <h1 align="center"><font size="5" face="arial" color="black">*****</font></h1>
  16. </body>
  17. </html>

下面进入本博客重点！！！

nx板代码，为后端（服务器）提供视频流，并叠加显示推送的检测信息。

  1. from flask import Flask, render_template, Response
  2. import asyncio
  3. import websockets
  4. import threading
  5. import pyzed.sl as sl
  6. import cv2
  7. import json
  8. import numpy as np
  9.  
  10. receive_dict = []
  11. receive_location = []
  12. receive_count = ""
  13. img = None
  14.  
  15.  
  16. init = sl.InitParameters()
  17. init.camera_resolution = sl.RESOLUTION.HD720
  18. init.depth_mode = sl.DEPTH_MODE.NONE
  19. cam = sl.Camera()
  20. status = cam.open(init)
  21. if status != sl.ERROR_CODE.SUCCESS:
  22.     print(repr(status))
  23.     exit(1)
  24.  
  25. runtime = sl.RuntimeParameters()
  26. mat = sl.Mat()
  27.  
  28. stream = sl.StreamingParameters()
  29. stream.codec = sl.STREAMING_CODEC.H264
  30. stream.bitrate = 4000
  31. status = cam.enable_streaming(stream)
  32. if status != sl.ERROR_CODE.SUCCESS:
  33.     print(repr(status))
  34.     exit(1)
  35.  
  36. print("  Quit : CTRL+C\n")
  37.  
  38. # Websocket服务端
  39. class WebsocketChatServer():
  40.     def __init__(self):
  41.         pass
  42.  
  43.     async def run(self, port):
  44.         start_server = websockets.serve(self.handler, "", port, ping_interval=None)
  45.         await start_server
  46.         print(f'  > server start ok! on port {port}')
  47.         await asyncio.Future()           # run forever
  48.  
  49.     async def handler(self, websocket, path):
  50.         global receive_dict, receive_location, receive_count
  51.         async for message in websocket:
  52.             receive_data = message
  53.             receive_count = json.loads(receive_data)["person_count"]
  54.             receive_dict = json.loads(receive_data)["detection_box"]
  55.             receive_location = json.loads(receive_data)["location"]
  56.             # print(receive_count)
  57.  
  58.  
  59. # 初始化和定义flask
  60. app = Flask(__name__)
  61.  
  62.  
  63. @app.route('/')
  64. def index():
  65.     test_dict = 'Welcome to use the impediment perception web service!'
  66.     return test_dict
  67.  
  68.  
  69. @app.route("/demo")
  70. def demo():
  71.     return render_template("index.html")
  72.  
  73.  
  74. @app.route("/video_feed")
  75. def video_feed():
  76.     return Response(generate(), mimetype="multipart/x-mixed-replace; boundary=frame")
  77.  
  78. # main函数主要实现向后端（服务器）推送视频流，并在视频流上叠加后端（服务器）推送的检测信息
  79. def main():
  80.     global img
  81.     while True:
  82.         err = cam.grab(runtime)
  83.         if (err == sl.ERROR_CODE.SUCCESS):
  84.             cam.retrieve_image(mat, sl.VIEW.LEFT)
  85.             # image = cv2.cvtColor(mat.get_data(), cv2.COLOR_RGB2BGR)
  86.             # image_cv = mat.get_data()
  87.             # image = image_cv[:, :, 0:3]
  88.  
  89.             image = mat.get_data()
  90.  
  91.             if len(receive_location) != 0 and len(receive_dict) != 0:
  92.                 if len(receive_location) != 1:
  93.                     last_data = receive_location[len(receive_location)-1]
  94.                     for i in range(len(last_data)):
  95.                         location_dict = last_data[i]
  96.                         # dict2str
  97.                         location_str = json.dumps(location_dict)
  98.                         # str2json
  99.                         location_json = json.loads(location_str)
  100.                         # print(location_json)
  101.                         if receive_count != 0:
  102.                             text = "(" + str(location_json['x']) + " " + str(location_json['y']) + " " + str(
  103.                                                 location_json['z']) + ")"
  104.                             cv2.circle(image, (int(location_json['x0']), int(location_json['y0'])), 2, (0, 0, 255),
  105.                                        thickness=3)
  106.                             cv2.putText(image, text, (int(location_json['x0']), int(location_json['y0'])),
  107.                                         cv2.FONT_HERSHEY_PLAIN, 2.0, (0, 0, 255), 2)
  108.  
  109.                 else:
  110.                     if len(receive_location[0]) != 0:
  111.                         # print(receive_location)
  112.                         location_str = json.dumps(receive_location[0][0])
  113.                         location_json = json.loads(location_str)
  114.                         if receive_count != 0:
  115.                             text = "(" + str(location_json['x']) + " " + str(location_json['y']) + " " + str(
  116.                                 location_json['z']) + ")"
  117.                             cv2.circle(image, (int(location_json['x0']), int(location_json['y0'])), 2, (0, 0, 255),
  118.                                        thickness=3)
  119.                             cv2.putText(image, text, (int(location_json['x0']), int(location_json['y0'])),
  120.                                         cv2.FONT_HERSHEY_PLAIN, 2.0, (0, 0, 255), 2)
  121.  
  122.                 if len(receive_dict) == 1:
  123.                     # print(receive_dict)
  124.                     if len(receive_dict[0]) != 0:
  125.                         box_str = json.dumps(receive_dict[0][0])
  126.                         box_json = json.loads(box_str)
  127.                         if receive_count != 0:
  128.                             cv2.putText(image, "person", (int(box_json["x1"]), int(box_json["y1"])-5),
  129.                                         cv2.FONT_HERSHEY_PLAIN, 2.0, (255, 0, 0), 2)
  130.                             img = cv2.rectangle(image, (int(box_json["x1"]), int(box_json["y1"])),
  131.                                                 (int(box_json["x2"]), int(box_json["y2"])), (255, 0, 0), 2)
  132.                         else:
  133.                             img = image
  134.  
  135.                         
  136.                         cv2.imshow("ZED", img)
  137.                         key = cv2.waitKey(1)
  138.  
  139.                 else:
  140.                     last_box = receive_dict[len(receive_dict) - 1]
  141.                     # print(last_box)
  142.                     for i in range(len(last_box)):
  143.                         # print(receive_dict[len(receive_dict) - 1])
  144.                         box_dict = last_box[i]
  145.                         # dict2str
  146.                         box_str = json.dumps(box_dict)
  147.                         # str2json
  148.                         box_json = json.loads(box_str)
  149.                         # print(box_json)
  150.                         if receive_count != 0:
  151.                             cv2.putText(image, "person", (int(box_json["x1"]), int(box_json["y1"]) - 5),
  152.                                         cv2.FONT_HERSHEY_PLAIN, 2.0, (255, 0, 0), 2)
  153.                             img = cv2.rectangle(image, (int(box_json["x1"]), int(box_json["y1"])),
  154.                                                 (int(box_json["x2"]), int(box_json["y2"])), (255, 0, 0), 2)
  155.                         else:
  156.                             img = image
  157.  
  158.                         
  159.                         cv2.imshow("ZED", img)
  160.                         key = cv2.waitKey(1)
  161.  
  162.             # else:
  163.             #     cv2.imshow("ZED", image)
  164.             #     key = cv2.waitKey(1)
  165.  
  166.         else:
  167.             key = cv2.waitKey(1)
  168.  
  169.     cam.disable_streaming()
  170.     cam.close()
  171.  
  172.  
  173. def generate():
  174.     # global img_result, lock
  175.     # loop over frames from the output stream
  176.     while True:
  177.         # wait until the lock is acquired
  178.         # with lock:
  179.         if img is None:
  180.             continue
  181.         # print("-----------------")
  182.         # cv2.imshow("ZED", img)
  183.         # key = cv2.waitKey(1)
  184.         (flag, encodedImage) = cv2.imencode(".jpg", img)
  185.         if not flag:
  186.             continue
  187.         # yield the output frame in the byte format
  188.         frame = encodedImage.tobytes()
  189.         # yield(b'--frame\r\n' b'Content-Type: image/jpeg\r\n\r\n' +
  190.         #       bytearray(encodedImage) + b'\r\n')
  191.         yield (b'--frame\r\n' b'Content-Type: image/jpeg\r\n\r\n' +
  192.                frame + b'\r\n')
  193.  
  194.  
  195. # 启动函数
  196. def main01():
  197.     # 在新线程中启动flask
  198.     # print(app.url_map)
  199.     # 启动websocket server
  200.     print('> starting server...')
  201.     server = WebsocketChatServer()
  202.     tasks = [
  203.         server.run(2333),
  204.     ]
  205.     # loop = asyncio.get_event_loop()
  206.     loop = asyncio.new_event_loop()
  207.     asyncio.set_event_loop(loop)
  208.     try:
  209.         loop.run_until_complete(asyncio.wait(tasks))
  210.     except KeyboardInterrupt:
  211.         for task in asyncio.Task.all_tasks():
  212.             task.cancel()
  213.         loop.stop()
  214.         loop.run_forever()
  215.     loop.close()
  216.  
  217.  
  218. if __name__ == '__main__':
  219.     flask_thread = threading.Thread(target=app.run, args=('0.0.0.0', 5000))
  220.     flask_thread.start()
  221.  
  222.     # socket_thread = threading.Thread(target=main)
  223.     # socket_thread.start()
  224.     threads = [threading.Thread(target=main), threading.Thread(target=main01)]
  225.     for t in threads:
  226.         t.start()
  227.     for t in threads:
  228.         t.join()
  229.

注：asyncio.wait()搜集所有任务;

运用线程，首先要生成一个loop对象，然后loop.run_xxx()就可以运行线程了，而如何创建这个loop, 方法有两种：1、对于主线程是loop=get_event_loop()；2、对于其他线程需要首先loop=new_event_loop()，然后set_event_loop(loop)。new_event_loop()是创建一个event loop对象，而set_event_loop(eventloop对象)是将event loop对象指定为当前线程的event loop。一个线程任务，不能运行在两个及两个以上不同的循环中，一个循环体可以运行多个不同的协程任务。

index.html

  1. <html>
  2.   <head>
  3.       <link rel="icon" href="{{ url_for('static', filename='favicon.ico') }}" type="image/x-icon">
  4.     <title>感知效果示意</title>
  5.     <style type="text/css">
  6.         .item .pic { float:left;margin-left:39%;width:30px;height:30px; }
  7.         .item .content {float:left;width:290px;}
  8.         .item .pic2 { float:left;margin-left:0px;width:30px;height:30px; }
  9.     </style>
  10.     <div class="item">
  11.         <div class="pic"><img src="{{ url_for('static', filename='favicon.ico') }}" width='30px' height='30px'></div>
  12.         <div class="content" style="font-size:22px; ">&ensp;感知效果&ensp; </div>
  13.         <div class="pic2"><img src="{{ url_for('static', filename='favicon.ico') }}" width='30px' height='30px'></div>
  14.     </div>
  15.       <br>
  16.       <br>
  17.   </head>
  18.   <body>
  19.     <div style="text-align: center;" >
  20.         <img src="{{url_for('video_feed')}}" width='1400px' height='800px'>
  21.     </div>
  22.     <div style="font-size:16px; text-align: center;">Copyright © ***</div>
  23.  
  24.   </body>
  25. </html>

 后端（服务器）代码与博客【ZED-2】基于WebSockets库，通过多线程方式推送数据流_WXG1011的博客-CSDN博客一致，这里不再粘贴。

参考链接 nohtypython | 同时使用flask和websockets - Mz1 - 博客园p

声明：本文内容由网友自发贡献，不代表【wpsshop博客】立场，版权归原作者所有，本站不承担相应法律责任。如您发现有侵权的内容，请联系我们。转载请注明出处：https://www.wpsshop.cn/w/凡人多烦事01/article/detail/280087
推荐阅读
相关标签

Copyright © 2003-2013 www.wpsshop.cn 版权所有，并保留所有权利。

  

闽ICP备14008679号