python:基于套接字socket传输视频的方法代码记录总结

最新推荐文章于 2024-07-09 15:52:26 发布

我爱写BUG

最新推荐文章于 2024-07-09 15:52:26 发布

阅读量4.5k

点赞数 8

分类专栏：深度学习与Python 文章标签： python socket 多线程 opencv TCP

本文链接：https://blog.csdn.net/qq_43797817/article/details/108382801

版权

深度学习与Python 专栏收录该内容

39 篇文章 5 订阅

订阅专栏

初始想法是想模拟IP摄像头传输视频然后实时监测的场景，后来发现和要实现的目标有所偏离，但还是把其中可参考、有所修改的可用代码记录下来。
功能都是客户端发送视频，服务器接收视频，没有服务器发送视频返回客户端的内容。

我在修改其中代码的时候没有考虑到多进程(multiprocessing)与多线程(threading)的区别,所以代码中有混用的情况，请注意这个情况，在此不做仔细区分。
下面多进程/多线程客户端的代码是相同的，可以多建几个文件只改变发送的视频文件，模拟多客户端。

先运行接收端（服务器），再运行发送端（客户端）。
ip地址要在同一网段下，及前三段数字相同，发送端被拒绝连接时要查看端口是否被占用。

1.使用opencv读取视频，按帧使用套接字发送并接收写为视频

import socket
import cv2
import numpy as np
import time
import sys

def SendVideo():
    address = ('172.23.13.216', 8025)  
    try:
        sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        sock.connect(address)
    except socket.error as msg:
        print(msg)
        sys.exit(1)

    capture = cv2.VideoCapture('../testvideo/video1_24.avi')
    # capture = cv2.VideoCapture(1)
    # capture.set(3, 1920)#参数：3：在视频流的帧的宽度  功能：把视频流的帧(图片)的宽度调成指定值
    # capture.set(4, 1080) #参数：4：在视频流的帧的高度  功能：把视频流的帧(图片)的高度调成指定值
    # capture.set(5, 24)                     #5:CV_CAP_PROP_FPS 帧速率

    ret, frame = capture.read()
    encode_param = [int(cv2.IMWRITE_JPEG_QUALITY), 15]
    sumtime = 0
    while ret:
        t1 = time.time()
        ## 首先对图片进行编码，因为socket不支持直接发送图片
        result, imgencode = cv2.imencode('.jpg', frame, encode_param)#
        data = np.array(imgencode)
        stringData = data.tostring()
        ## 首先发送图片编码后的长度
        sock.send(str.encode(str(len(stringData)).ljust(16)))
        ## 然后发送编码的内容
        sock.send(stringData)

        eachtime = time.time() - t1
        sumtime += eachtime
        ret, frame = capture.read()# cv2.imshow('show', frame)
        if cv2.waitKey(
                2) & 0xff == 27:
            break

        print('process time = ', eachtime)  # 处理时间

    sock.close()

    print('++++++++++This path send over++++++++++')
    print(sumtime)

if __name__ == '__main__':
    SendVideo()

"""
传输时间慢
1.cv2.imencode('.jpg', frame, encode_param),修改为png,传输速度更慢
2.encode_param = [int(cv2.IMWRITE_JPEG_QUALITY), 15]，压缩传输中的质量参数改变对传输速度影响小
原因 待寻找
"""

import socket
import time
import cv2
import numpy
import copy
from multiprocessing import Process#多进程


def ReceiveVideo(conn, addr):
    # 接受图片及大小的信息
    def recvall(sock, count):#读取count长度的数据
        buf = b''
        while count:
            newbuf = sock.recv(count)  # s.sendall()发送, s.recv()接收. 因为client是通过 sk.recv()来进行接受数据，而count表示，最多每次接受count字节，
            if not newbuf: return None
            buf += newbuf
            count -= len(newbuf)
        return buf

    print('connect from:' + str(addr))
    fourcc = cv2.VideoWriter_fourcc('X', 'V', 'I','D')#avi格式
    videoWriter = cv2.VideoWriter('./receive_video/%s.avi' % time.strftime("%Y-%m-%d-%H_%M_%S",time.localtime(time.time())) , fourcc, 24, (1920, 1080))##%(str(time.strftime("%Y-%m-%d-%H:%M:%S",time.localtime())))
    while True:
        start = time.time()
        length = recvall(conn, 1024)
        stringData = recvall(conn, int(length))#nt()只能转化由纯数字组成的字符串

        # length =conn.recv(1024)
        # stringData = conn.recv(int(length))#nt()只能转化由纯数字组成的字符串


        data = numpy.fromstring(stringData,numpy.uint8)  # 将获取到的字符流数据转换成1维数组 data = numpy.fromstring()           numpy.frombuffer
        decimg = cv2.imdecode(data, cv2.IMREAD_COLOR)  # 将数组解码成图像
        #cv2.imshow('SERVER', decimg)  # 显示图像
        # =================================================================================================================================
        videoWriter.write(decimg)
        end = time.time()
        eachtime =end - start
        print('process time = ', eachtime)
        # ================================================================================================================================
        if cv2.waitKey(
                10) & 0xff == 27:
            break
    cv2.destroyAllWindows()
    # videoWriter.release()##释放写入视频流


if __name__ == '__main__':
    address = ('172.23.13.226', 8035)#'192.168.1.104', 8004
    s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    s.bind(address)
    s.listen(5)
    while True:
        conn, addr = s.accept()
        p=Process(target=ReceiveVideo,args=(conn, addr))  #daemon默认值为False，如果设置为True，代表该进程为后台守护进程；当该进程的父进程终止时，该进程也随之终止；并且设置为True后，该进程不能创建子进程，设置该属性必须在start()之前
        p.start()
    s.close()

2.读写文件形式，二进制文件，直接发送及接收为视频

参考：Python socket传输MP4,MP3

import socket
import os
from time import *
import sys


def SendVideo():
    address = ('172.23.13.226', 8105)  # 192.168.1.103ubuntu
    try:
        sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        sock.connect(address)
    except socket.error as msg:
        print(msg)
        sys.exit(1)
    videopath = "../testvideo/video0_24.flv"
    video_size = os.path.getsize(videopath)/1024/1024  #字节换算为MB
    with open(videopath,'rb') as f:
        stringData = f.read() #读出来就是字节

    start_time = time()
    sock.send(stringData)
    end_time = time()

    All_time = end_time - start_time
    rate = round(video_size/All_time,2) #计算速率，四舍五入保留两位
    print("[*]发送时间(sec)：%s"%(All_time))
    print(f"[*]Rate： {rate}MB/s")
    print('[*]This path send over')
    sock.close()
if __name__ == '__main__':
    SendVideo()

import socket
import time
import os
from multiprocessing import Process

def ReceiveVideo(conn,addr):
    print('[*]connect from:' + str(addr))
    start = time.time()
    videoname="%s.flv" % (time.strftime("%Y-%m-%d-%H_%M_%S",time.localtime(time.time())))

    while True:#一次接收1024字节 持续发送
        s=time.time()
        stringData=conn.recv(1024)
        if not stringData: break
        videopath="./receive_video/"+ videoname          #% time.strftime("%Y-%m-%d-%H_%M_%S",time.localtime(time.time()))
        with open(videopath,'ab') as f:
            f.write(stringData)
        #=====================
        # e = time.time()
        # eachtime=e-s
        # print(eachtime)
        # each_size=os.path.getsize(videopath)
        # print(each_size)
        # rate = round(each_size / eachtime, 4)
        # print(print(f"[*]Rate： {rate}MB/s"))
        #=======================

    end = time.time()
    print('[*]process time (sec): ', end - start)
    video_size = os.path.getsize(videopath)# #单位是B(字节)
    print(f"[*]This File is {round(video_size/1024/1024,2)} MB")
    print('[*]This path send over')

if __name__ == '__main__':
    address = ('172.23.13.226', 8105)#'192.168.1.104', 8004
    s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    s.bind(address)
    s.listen(5)
    print(f"[*]Listening:  {address}")
    while True:
        conn, addr = s.accept()
        p=Process(target=ReceiveVideo,args=(conn, addr))  #daemon默认值为False，如果设置为True，代表该进程为后台守护进程；当该进程的父进程终止时，该进程也随之终止；并且设置为True后，该进程不能创建子进程，设置该属性必须在start()之前
        p.start()
    s.close()

3.

参考：
Python3远程监控程序实现
 基于python和opencv的视频传输程序(一)

import socket
import struct
import time
import cv2
import numpy


class Config(object):
    def __init__(self):
        self.TargetIP = ('172.23.13.216', 6666)
        self.resolution = (720, 480)  # 分辨率720, 480
        self.img_fps = 15  # each second send pictures
        self.server = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        self.server.connect(self.TargetIP)
        self.img = ''
        self.img_data = ''

    def RT_Image(self):
        camera = cv2.VideoCapture(0)#"../testvideo/video0_24.flv"
        img_param = [int(cv2.IMWRITE_JPEG_QUALITY), self.img_fps]

        while True:
            #time.sleep(0.1)  # sleep for 0.1 seconds
            ret, self.img = camera.read()
            self.img = cv2.resize(self.img, self.resolution)

            _, img_encode = cv2.imencode('.jpg', self.img, img_param)
            img_code = numpy.array(img_encode)
            self.img_data = img_code.tostring()  # bytes data
            try:

                packet = struct.pack(b'lhh', len(self.img_data), self.resolution[0],
                                     self.resolution[1])
                self.server.send(packet)
                self.server.send(self.img_data)

            except Exception as e:
                print(e.args)
                camera.release()
                return


if __name__ == '__main__':
    config = Config()
    config.RT_Image()

import socket
import cv2
import struct
import numpy
import threading
import time


class Camera_Connect_Object(object):
    def __init__(self, TargetIP=('', 6666)):
        self.TargetIP = TargetIP
        self.resolution = (720, 480)
        self.src = 888 + 15
        self.interval = 0
        self.img_fps = 15

        self.Server = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        self.Server.bind(self.TargetIP)
        self.Server.listen(5)

    def RT_Image(self):

        #===========================================================================================================
        fourcc = cv2.VideoWriter_fourcc('X', 'V', 'I', 'D')  # avi格式
        videoWriter = cv2.VideoWriter('./%s.avi' % time.strftime("%Y-%m-%d-%H_%M_%S", time.localtime(time.time())),
                                      fourcc, 30, (720, 480))
        # ===========================================================================================================

        self.client, self.addr = self.Server.accept()
        self.name = self.addr[0] + " Camera"
        print(self.name)
        while True:
            # time.sleep(0.3)  # sleep for 0.3 seconds
            tempdata = self.client.recv(8)
            if len(tempdata) == 0:
                print("+1")
                continue
            info = struct.unpack('lhh', tempdata)
            buf_size = int(info[0])

            if buf_size:
                try:
                    self.buf = b""
                    self.temp_buf = self.buf
                    while buf_size:
                        self.temp_buf = self.client.recv(buf_size)
                        buf_size -= len(self.temp_buf)
                        self.buf += self.temp_buf
                    data = numpy.fromstring(self.buf, dtype='uint8')
                    self.image = cv2.imdecode(data, 1)
                    # ===========================================================================================================
                    videoWriter.write(self.image)
                    # ===========================================================================================================
                    cv2.imshow(self.name, self.image)
                except Exception as e:
                    print(e.args)
                    pass
                finally:
                    if cv2.waitKey(10) == 27:
                        self.client.close()
                        cv2.destroyAllWindows()

                        videoWriter.release()

                        break

    def Get_data(self):
        showThread = threading.Thread(target=self.RT_Image)
        showThread.start()
        showThread.join()


if __name__ == '__main__':
    camera = Camera_Connect_Object()
    camera.Get_data()

在个我在自己windows电脑本地发送接收成功，但在windoes向ubuntu发送接收失败，ubuntu出现有关于“info = struct.unpack(‘lhh’, tempdata)”的错误提示。

4.

参考
基于Python3 + OpenCV3.3.1的远程监控程序

#服务器端

import socket
import threading
import struct
import time
import cv2
import numpy

class Carame_Accept_Object:
    def __init__(self,S_addr_port=("",8880)):
        self.resolution=(640,480)       #分辨率
        self.img_fps=15                 #每秒传输多少帧数
        self.addr_port=S_addr_port
        self.Set_Socket(self.addr_port)

    #设置套接字
    def Set_Socket(self,S_addr_port):
        self.server=socket.socket(socket.AF_INET,socket.SOCK_STREAM)
        self.server.setsockopt(socket.SOL_SOCKET,socket.SO_REUSEADDR,1) #端口可复用
        self.server.bind(S_addr_port)
        self.server.listen(5)
        #print("the process work in the port:%d" % S_addr_port[1])


def check_option(object,client):
    #按格式解码，确定帧数和分辨率
    info=struct.unpack('lhh',client.recv(8))
    if info[0]>888:
        object.img_fps=int(info[0])-888          #获取帧数
        object.resolution=list(object.resolution)
        # 获取分辨率
        object.resolution[0]=info[1]
        object.resolution[1]=info[2]
        object.resolution = tuple(object.resolution)
        return 1
    else:
        return 0

def RT_Image(object,client,D_addr):
    if(check_option(object,client)==0):
        return
    camera=cv2.VideoCapture(0)                                #从摄像头中获取视频
    img_param=[int(cv2.IMWRITE_JPEG_QUALITY),object.img_fps]  #设置传送图像格式、帧数
    while(1):
        time.sleep(0.1)             #推迟线程运行0.1s
        _,object.img=camera.read()  #读取视频每一帧

        object.img=cv2.resize(object.img,object.resolution)     #按要求调整图像大小(resolution必须为元组)
        _,img_encode=cv2.imencode('.jpg',object.img,img_param)  #按格式生成图片
        img_code=numpy.array(img_encode)                        #转换成矩阵
        object.img_data=img_code.tostring()                     #生成相应的字符串
        try:
            #按照相应的格式进行打包发送图片
            client.send(struct.pack("lhh",len(object.img_data),object.resolution[0],object.resolution[1])+object.img_data)
        except:
            camera.release()        #释放资源
            return

if __name__ == '__main__':
    camera=Carame_Accept_Object()
    while(1):
        client,D_addr=camera.server.accept()
        clientThread=threading.Thread(None,target=RT_Image,args=(camera,client,D_addr,))
        clientThread.start()

#客户端


import socket
import cv2
import threading
import struct
import numpy

class Camera_Connect_Object:
    def __init__(self,D_addr_port=["",8880]):
        self.resolution=[640,480]
        self.addr_port=D_addr_port
        self.src=888+15                 #双方确定传输帧数，（888）为校验值
        self.interval=0                 #图片播放时间间隔
        self.img_fps=15                 #每秒传输多少帧数

    def Set_socket(self):
        self.client=socket.socket(socket.AF_INET,socket.SOCK_STREAM)
        self.client.setsockopt(socket.SOL_SOCKET,socket.SO_REUSEADDR,1)

    def Socket_Connect(self):
        self.Set_socket()
        self.client.connect(self.addr_port)
        print("IP is %s:%d" % (self.addr_port[0],self.addr_port[1]))

    def RT_Image(self):
        #按照格式打包发送帧数和分辨率
        self.name=self.addr_port[0]+" Camera"
        self.client.send(struct.pack("lhh", self.src, self.resolution[0], self.resolution[1]))
        while(1):
            info=struct.unpack("lhh",self.client.recv(8))
            buf_size=info[0]                    #获取读的图片总长度
            if buf_size:
                try:
                    self.buf=b""                #代表bytes类型
                    temp_buf=self.buf
                    while(buf_size):            #读取每一张图片的长度
                        temp_buf=self.client.recv(buf_size)
                        buf_size-=len(temp_buf)
                        self.buf+=temp_buf      #获取图片
                        data = numpy.fromstring(self.buf, dtype='uint8')    #按uint8转换为图像矩阵
                        self.image = cv2.imdecode(data, 1)                  #图像解码
                        cv2.imshow(self.name, self.image)                   #展示图片
                except:
                    pass;
                finally:
                    if(cv2.waitKey(10)==27):        #每10ms刷新一次图片，按‘ESC’（27）退出
                        self.client.close()
                        cv2.destroyAllWindows()
                        break

    def Get_Data(self,interval):
        showThread=threading.Thread(target=self.RT_Image)
        showThread.start()

if __name__ == '__main__':
    camera=Camera_Connect_Object()
    camera.addr_port[0]=input("Please input IP:")
    camera.addr_port=tuple(camera.addr_port)
    camera.Socket_Connect()
    camera.Get_Data(camera.interval)