我正在尝试使用 Python 从 IP 摄像机流式传输 RTP 数据包。
我可以使用 RTSP 协议(protocol)发送描述、设置和播放命令,但是,我无法使用 RTP 开始流式传输实际视频流。
代码如下:
import socket
def printrec(recst):
recs=recst.split('\r\n')
for rec in recs:
print rec
dest="DESCRIBE rtsp://admin:12345@192.168.1.74 RTSP/1.0\r\nCSeq: 2\r\nUser-Agent: python\r\nAccept: application/sdp\r\n\r\n"
setu="SETUP rtsp://admin:12345@192.168.1.74/trackID=1 RTSP/1.0\r\nCSeq: 3\r\nUser-Agent: python\r\nTransport: RTP/AVP;unicast;client_port=60784-60785\r\n\r\n"
play="PLAY rtsp://admin:12345@192.168.1.74/ RTSP/1.0\r\nCSeq: 5\r\nUser-Agent: python\r\nSession: SESID\r\nRange: npt=0.000-\r\n\r\n"
# .. here SESID will be substituted with the session id that SETUP returns us ..
ip="192.168.1.74"
s=socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.connect((ip,554))
s.send(dest)
recst=s.recv(4096)
printrec(recst)
# etc. then the same for the strins "setu" and "play"
....
安装程序回复我:
RTSP/1.0 200 OK
CSeq: 3
Session: 1117448817;timeout=60
Transport: RTP/AVP;unicast;client_port=60784-60785;server_port=8214-8215;ssrc=40d35c30;mode="play"
Date: Mon, Jan 19 2015 11:10:08 GMT
所以我们有客户端端口 60784 和 60785 以及服务器端口 8214 和 8215。
接下来,play回复我:
RTSP/1.0 200 OK
CSeq: 5
Session: 1117448817
RTP-Info: url=rtsp://admin:12345@192.168.1.74/trackID=1;seq=3539; rtptime=16026930
Date: Mon, Jan 19 2015 11:10:08 GMT
接下来,我应该怎么做才能开始接收 RTP 数据包? 打开一个 UDP 套接字如下.. ?
sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM) # UDP
sock.bind((ip, serverport))
sock.recv(4096)
但我什么也没得到 ..(它停止了)
请原谅我对套接字的了解不多..!
最佳答案
所以,
经过大量的谷歌搜索和 wireshark 分析后,我想出了正确的解决方案。
我在这里发布生成的演示代码。我认为它可能对社区有用。
如果你想用 python 读取 ip-cams 并将 H264 流转储到可食用文件中,这就是你要找的东西。
尽情享受吧!
"""
A demo python code that ..
1) Connects to an IP cam with RTSP
2) Draws RTP/NAL/H264 packets from the camera
3) Writes them to a file that can be read with any stock video player (say, mplayer, vlc & other ffmpeg based video-players)
Done for educative/demonstrative purposes, not for efficiency..!
written 2015 by Sampsa Riikonen.
"""
import socket
import re
import bitstring # if you don't have this from your linux distro, install with "pip install bitstring"
# ************************ FOR QUICK-TESTING EDIT THIS AREA *********************************************************
ip="192.168.1.74" # IP address of your cam
adr="rtsp://admin:12345@192.168.1.74" # username, passwd, etc.
clientports=[60784,60785] # the client ports we are going to use for receiving video
fname="stream.h264" # filename for dumping the stream
rn=5000 # receive this many packets
# After running this program, you can try your file defined in fname with "vlc fname" or "mplayer fname" from the command line
# you might also want to install h264bitstream to analyze your h264 file
# *******************************************************************************************************************
dest="DESCRIBE "+adr+" RTSP/1.0\r\nCSeq: 2\r\nUser-Agent: python\r\nAccept: application/sdp\r\n\r\n"
setu="SETUP "+adr+"/trackID=1 RTSP/1.0\r\nCSeq: 3\r\nUser-Agent: python\r\nTransport: RTP/AVP;unicast;client_port="+str(clientports[0])+"-"+str(clientports[1])+"\r\n\r\n"
play="PLAY "+adr+" RTSP/1.0\r\nCSeq: 5\r\nUser-Agent: python\r\nSession: SESID\r\nRange: npt=0.000-\r\n\r\n"
# File organized as follows:
# 1) Strings manipulation routines
# 2) RTP stream handling routine
# 3) Main program
# *** (1) First, some string searching/manipulation for handling the rtsp strings ***
def getPorts(searchst,st):
""" Searching port numbers from rtsp strings using regular expressions
"""
pat=re.compile(searchst+"=\d*-\d*")
pat2=re.compile('\d+')
mstring=pat.findall(st)[0] # matched string .. "client_port=1000-1001"
nums=pat2.findall(mstring)
numas=[]
for num in nums:
numas.append(int(num))
return numas
def getLength(st):
""" Searching "content-length" from rtsp strings using regular expressions
"""
pat=re.compile("Content-Length: \d*")
pat2=re.compile('\d+')
mstring=pat.findall(st)[0] # matched string.. "Content-Length: 614"
num=int(pat2.findall(mstring)[0])
return num
def printrec(recst):
""" Pretty-printing rtsp strings
"""
recs=recst.split('\r\n')
for rec in recs:
print rec
def sessionid(recst):
""" Search session id from rtsp strings
"""
recs=recst.split('\r\n')
for rec in recs:
ss=rec.split()
# print ">",ss
if (ss[0].strip()=="Session:"):
return int(ss[1].split(";")[0].strip())
def setsesid(recst,idn):
""" Sets session id in an rtsp string
"""
return recst.replace("SESID",str(idn))
# ********* (2) The routine for handling the RTP stream ***********
def digestpacket(st):
""" This routine takes a UDP packet, i.e. a string of bytes and ..
(a) strips off the RTP header
(b) adds NAL "stamps" to the packets, so that they are recognized as NAL's
(c) Concantenates frames
(d) Returns a packet that can be written to disk as such and that is recognized by stock media players as h264 stream
"""
startbytes="\x00\x00\x00\x01" # this is the sequence of four bytes that identifies a NAL packet.. must be in front of every NAL packet.
bt=bitstring.BitArray(bytes=st) # turn the whole string-of-bytes packet into a string of bits. Very unefficient, but hey, this is only for demoing.
lc=12 # bytecounter
bc=12*8 # bitcounter
version=bt[0:2].uint # version
p=bt[3] # P
x=bt[4] # X
cc=bt[4:8].uint # CC
m=bt[9] # M
pt=bt[9:16].uint # PT
sn=bt[16:32].uint # sequence number
timestamp=bt[32:64].uint # timestamp
ssrc=bt[64:96].uint # ssrc identifier
# The header format can be found from:
# https://en.wikipedia.org/wiki/Real-time_Transport_Protocol
lc=12 # so, we have red twelve bytes
bc=12*8 # .. and that many bits
print "version, p, x, cc, m, pt",version,p,x,cc,m,pt
print "sequence number, timestamp",sn,timestamp
print "sync. source identifier",ssrc
# st=f.read(4*cc) # csrc identifiers, 32 bits (4 bytes) each
cids=[]
for i in range(cc):
cids.append(bt[bc:bc+32].uint)
bc+=32; lc+=4;
print "csrc identifiers:",cids
if (x):
# this section haven't been tested.. might fail
hid=bt[bc:bc+16].uint
bc+=16; lc+=2;
hlen=bt[bc:bc+16].uint
bc+=16; lc+=2;
print "ext. header id, header len",hid,hlen
hst=bt[bc:bc+32*hlen]
bc+=32*hlen; lc+=4*hlen;
# OK, now we enter the NAL packet, as described here:
#
# https://tools.ietf.org/html/rfc6184#section-1.3
#
# Some quotes from that document:
#
"""
5.3. NAL Unit Header Usage
The structure and semantics of the NAL unit header were introduced in
Section 1.3. For convenience, the format of the NAL unit header is
reprinted below:
+---------------+
|0|1|2|3|4|5|6|7|
+-+-+-+-+-+-+-+-+
|F|NRI| Type |
+---------------+
This section specifies the semantics of F and NRI according to this
specification.
"""
"""
Table 3. Summary of allowed NAL unit types for each packetization
mode (yes = allowed, no = disallowed, ig = ignore)
Payload Packet Single NAL Non-Interleaved Interleaved
Type Type Unit Mode Mode Mode
-------------------------------------------------------------
0 reserved ig ig ig
1-23 NAL unit yes yes no
24 STAP-A no yes no
25 STAP-B no no yes
26 MTAP16 no no yes
27 MTAP24 no no yes
28 FU-A no yes yes
29 FU-B no no yes
30-31 reserved ig ig ig
"""
# This was also very usefull:
# http://stackoverflow.com/questions/7665217/how-to-process-raw-udp-packets-so-that-they-can-be-decoded-by-a-decoder-filter-i
# A quote from that:
"""
First byte: [ 3 NAL UNIT BITS | 5 FRAGMENT TYPE BITS]
Second byte: [ START BIT | RESERVED BIT | END BIT | 5 NAL UNIT BITS]
Other bytes: [... VIDEO FRAGMENT DATA...]
"""
fb=bt[bc] # i.e. "F"
nri=bt[bc+1:bc+3].uint # "NRI"
nlu0=bt[bc:bc+3] # "3 NAL UNIT BITS" (i.e. [F | NRI])
typ=bt[bc+3:bc+8].uint # "Type"
print "F, NRI, Type :", fb, nri, typ
print "first three bits together :",bt[bc:bc+3]
if (typ==7 or typ==8):
# this means we have either an SPS or a PPS packet
# they have the meta-info about resolution, etc.
# more reading for example here:
# http://www.cardinalpeak.com/blog/the-h-264-sequence-parameter-set/
if (typ==7):
print ">>>>> SPS packet"
else:
print ">>>>> PPS packet"
return startbytes+st[lc:]
# .. notice here that we include the NAL starting sequence "startbytes" and the "First byte"
bc+=8; lc+=1; # let's go to "Second byte"
# ********* WE ARE AT THE "Second byte" ************
# The "Type" here is most likely 28, i.e. "FU-A"
start=bt[bc] # start bit
end=bt[bc+2] # end bit
nlu1=bt[bc+3:bc+8] # 5 nal unit bits
if (start): # OK, this is a first fragment in a movie frame
print ">>> first fragment found"
nlu=nlu0+nlu1 # Create "[3 NAL UNIT BITS | 5 NAL UNIT BITS]"
head=startbytes+nlu.bytes # .. add the NAL starting sequence
lc+=1 # We skip the "Second byte"
if (start==False and end==False): # intermediate fragment in a sequence, just dump "VIDEO FRAGMENT DATA"
head=""
lc+=1 # We skip the "Second byte"
elif (end==True): # last fragment in a sequence, just dump "VIDEO FRAGMENT DATA"
head=""
print "<<<< last fragment found"
lc+=1 # We skip the "Second byte"
if (typ==28): # This code only handles "Type" = 28, i.e. "FU-A"
return head+st[lc:]
else:
raise(Exception,"unknown frame type for this piece of s***")
# *********** (3) THE MAIN PROGRAM STARTS HERE ****************
# Create an TCP socket for RTSP communication
# further reading:
# https://docs.python.org/2.7/howto/sockets.html
s=socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.connect((ip,554)) # RTSP should peek out from port 554
print
print "*** SENDING DESCRIBE ***"
print
s.send(dest)
recst=s.recv(4096)
print
print "*** GOT ****"
print
printrec(recst)
print
print "*** SENDING SETUP ***"
print
s.send(setu)
recst=s.recv(4096)
print
print "*** GOT ****"
print
printrec(recst)
idn=sessionid(recst)
serverports=getPorts("server_port",recst)
clientports=getPorts("client_port",recst)
print "****"
print "ip,serverports",ip,serverports
print "****"
s1=socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
s1.bind(("", clientports[0])) # we open a port that is visible to the whole internet (the empty string "" takes care of that)
s1.settimeout(5) # if the socket is dead for 5 s., its thrown into trash
# further reading:
# https://wiki.python.org/moin/UdpCommunication
# Now our port is open for receiving shitloads of videodata. Give the camera the PLAY command..
print
print "*** SENDING PLAY ***"
print
play=setsesid(play,idn)
s.send(play)
recst=s.recv(4096)
print
print "*** GOT ****"
print
printrec(recst)
print
print
print "** STRIPPING RTP INFO AND DUMPING INTO FILE **"
f=open(fname,'w')
for i in range(rn):
print
print
recst=s1.recv(4096)
print "read",len(recst),"bytes"
st=digestpacket(recst)
print "dumping",len(st),"bytes"
f.write(st)
f.close()
# Before closing the sockets, we should give the "TEARDOWN" command via RTSP, but I am feeling lazy today (after googling, wireshark-analyzing, among other-things).
s.close()
s1.close()
关于python - 在 RTSP 设置后接收 RTP 数据包,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/28022432/
我有一个Ruby程序,它使用rubyzip压缩XML文件的目录树。gem。我的问题是文件开始变得很重,我想提高压缩级别,因为压缩时间不是问题。我在rubyzipdocumentation中找不到一种为创建的ZIP文件指定压缩级别的方法。有人知道如何更改此设置吗?是否有另一个允许指定压缩级别的Ruby库? 最佳答案 这是我通过查看rubyzip内部创建的代码。level=Zlib::BEST_COMPRESSIONZip::ZipOutputStream.open(zip_file)do|zip|Dir.glob("**/*")d
关闭。这个问题是opinion-based.它目前不接受答案。想要改进这个问题?更新问题,以便editingthispost可以用事实和引用来回答它.关闭4年前。Improvethisquestion我想在固定时间创建一系列低音和高音调的哔哔声。例如:在150毫秒时发出高音调的蜂鸣声在151毫秒时发出低音调的蜂鸣声200毫秒时发出低音调的蜂鸣声250毫秒的高音调蜂鸣声有没有办法在Ruby或Python中做到这一点?我真的不在乎输出编码是什么(.wav、.mp3、.ogg等等),但我确实想创建一个输出文件。
我在使用omniauth/openid时遇到了一些麻烦。在尝试进行身份验证时,我在日志中发现了这一点:OpenID::FetchingError:Errorfetchinghttps://www.google.com/accounts/o8/.well-known/host-meta?hd=profiles.google.com%2Fmy_username:undefinedmethod`io'fornil:NilClass重要的是undefinedmethodio'fornil:NilClass来自openid/fetchers.rb,在下面的代码片段中:moduleNetclass
我主要使用Ruby来执行此操作,但到目前为止我的攻击计划如下:使用gemsrdf、rdf-rdfa和rdf-microdata或mida来解析给定任何URI的数据。我认为最好映射到像schema.org这样的统一模式,例如使用这个yaml文件,它试图描述数据词汇表和opengraph到schema.org之间的转换:#SchemaXtoschema.orgconversion#data-vocabularyDV:name:namestreet-address:streetAddressregion:addressRegionlocality:addressLocalityphoto:i
我正在查看instance_variable_set的文档并看到给出的示例代码是这样做的:obj.instance_variable_set(:@instnc_var,"valuefortheinstancevariable")然后允许您在类的任何实例方法中以@instnc_var的形式访问该变量。我想知道为什么在@instnc_var之前需要一个冒号:。冒号有什么作用? 最佳答案 我的第一直觉是告诉你不要使用instance_variable_set除非你真的知道你用它做什么。它本质上是一种元编程工具或绕过实例变量可见性的黑客攻击
我想设置一个默认日期,例如实际日期,我该如何设置?还有如何在组合框中设置默认值顺便问一下,date_field_tag和date_field之间有什么区别? 最佳答案 试试这个:将默认日期作为第二个参数传递。youcorrectlysetthedefaultvalueofcomboboxasshowninyourquestion. 关于ruby-on-rails-date_field_tag,如何设置默认日期?[rails上的ruby],我们在StackOverflow上找到一个类似的问
有时我需要处理键/值数据。我不喜欢使用数组,因为它们在大小上没有限制(很容易不小心添加超过2个项目,而且您最终需要稍后验证大小)。此外,0和1的索引变成了魔数(MagicNumber),并且在传达含义方面做得很差(“当我说0时,我的意思是head...”)。散列也不合适,因为可能会不小心添加额外的条目。我写了下面的类来解决这个问题:classPairattr_accessor:head,:taildefinitialize(h,t)@head,@tail=h,tendend它工作得很好并且解决了问题,但我很想知道:Ruby标准库是否已经带有这样一个类? 最佳
我正在处理旧代码的一部分。beforedoallow_any_instance_of(SportRateManager).toreceive(:create).and_return(true)endRubocop错误如下:Avoidstubbingusing'allow_any_instance_of'我读到了RuboCop::RSpec:AnyInstance我试着像下面那样改变它。由此beforedoallow_any_instance_of(SportRateManager).toreceive(:create).and_return(true)end对此:let(:sport_
我正在玩HTML5视频并且在ERB中有以下片段:mp4视频从在我的开发环境中运行的服务器很好地流式传输到chrome。然而firefox显示带有海报图像的视频播放器,但带有一个大X。问题似乎是mongrel不确定ogv扩展的mime类型,并且只返回text/plain,如curl所示:$curl-Ihttp://0.0.0.0:3000/pr6.ogvHTTP/1.1200OKConnection:closeDate:Mon,19Apr201012:33:50GMTLast-Modified:Sun,18Apr201012:46:07GMTContent-Type:text/plain
这个问题在这里已经有了答案:关闭10年前。PossibleDuplicate:Pythonconditionalassignmentoperator对于这样一个简单的问题表示歉意,但是谷歌搜索||=并不是很有帮助;)Python中是否有与Ruby和Perl中的||=语句等效的语句?例如:foo="hey"foo||="what"#assignfooifit'sundefined#fooisstill"hey"bar||="yeah"#baris"yeah"另外,类似这样的东西的通用术语是什么?条件分配是我的第一个猜测,但Wikipediapage跟我想的不太一样。