Recherche avancée

Sur d’autres sites (13183)

• Running a py script in the Cloud

  12 janvier 2018, par Anay Bose

  I’m new to Google’s cloud & Virtual Machine(VM) instances, and I need some clarifications on a couple of points. I have a python script ; it imports a long range of functions. I need to run those functions in parallel. I’m using multiprocessing and Process, not threads. These functions are basically image and media processors, and they use many other tools like FFMPEG, imagemagick and Avisynth in addition to a wide range of python modules, including moviepy. Now, I would like to run some 50 functions in parallel assigning a CPU for each process. Images, media and avi files are stored in seperate folders. I’m on Windows7 Core-i7 machine. So, need cloud computing power.

  Now, my question can I run such a python script/app in the cloud that requires a very complicated file system and non-python tools i.e. ffmpeg, avisynth and avi files ?

  Can Google VMs emulate my local machine and empower me with more cores and memory to run such a program ? if not, then what are my options ? Is their any tutorials that I can follow ? I need your suggestions. I have given below an example script and some codes to help facilitate your understanding about my situation.

  from __future__ import unicode_literals
  
  import youtube_dl
  
  import os
  
  import time
  
  import sys
  
  reload(sys)  
  
  sys.setdefaultencoding('utf-8')
  
  from multiprocessing import Process
  
  from utils import *
  
  
  
  from clip31 import VIDEO31
  
  from clip32 import VIDEO32
  
  from clip189 import VIDEO189
  
  from clip16 import VIDEO16
  
  from clip39 import VIDEO39
  
  
  
  
  
  if __name__== '__main__':
  
  
  
      # 1. CALLING A FUNCTION
  
      folder = "bodyforce3\\16"
  
      serial = "16"
  
      images = get_filepaths("../16")
  
      videos = get_filepaths("12__media")
  
      pngs = get_filepaths("../pngs")
  
  
  
      Process(target=VIDEO192, args=(folder, serial, color1, color2, color3, images, videos)).start()
  
  
  
  
  
      # 2. CALLING A FUNCTION
  
      folder = "bodyforce3\\20"
  
      serial = "20"
  
      images = get_filepaths("../20")
  
      videos = get_filepaths("18__media")
  
  
  
      Process(target=VIDEO32, args=(folder, serial, color1, color2, color3, images, videos)).start()
  
  
  
  
  
      # 3. CALLING A FUNCTION
  
      folder = "bodyforce3\\14"
  
      serial = "14"
  
      images = get_filepaths("../14")
  
      videos = get_filepaths("16__media")
  
  
  
      Process(target=VIDEO91, args=(folder, serial, color1, color2, color3, images, videos)).start()

  I copy avi files in functions like this :

  src = "clip50_files"
  
  src_files = os.listdir(src)
  
  for file_name in src_files:
  
      full_file_name = os.path.join(src, file_name)
  
      if (os.path.isfile(full_file_name)):
  
          shutil.copy(full_file_name, folder)

  I call ffmpeg commands like this, and they are included within py functions.

  ###########################
  
  #### FFMPEG OPERATIONS ####
  
  ###########################
  
  
  
  print "Starting FFMPEG operations ..."
  
  
  
  if os.path.isfile(os.path.join(folder, "bounce-(3).avi")):
  
      os.remove(os.path.join(folder, "bounce-(3).avi"))
  
  
  
  
  
  infile = folder + "/bounce-(3).avs"
  
  outfile = folder + "/bounce-(3).avi"
  
  codec = "rawvideo"
  
  pix_fmt = "bgra"
  
  
  
  try:
  
      subprocess.call(["ffmpeg",
  
                       "-i" ,infile, 
  
                       "-c:v" ,codec, 
  
                       "-pix_fmt", pix_fmt, 
  
                       outfile],
  
                      stdout=open(os.devnull, 'w'), 
  
                      stderr=subprocess.STDOUT)
  
  except subprocess.CalledProcessError as e:   
  
      #except subprocess.CalledProcessError as e:
  
      sys.exit(e.output)
  
  except OSError as e:
  
      sys.exit(e.strerror)
  
  
  
  
  
  print "FFMPEG operations ended"

• FFMpeg access AVFoundation usb subdevice camera on OSX Mojave

  20 août 2020, par Retiarius

  I Have a dual USB camera for VR : two cameras, one usb connection. On linux, this appears in /dev/video0 and /dev/video1 and I can capture using ffmpeg -i /dev/video0

  



  On Mojave, I can see both devices in the USB hub :

  



  USB 2.0 Hub:

Product ID: 0x0101
Vendor ID:  0x1a40  (TERMINUS TECHNOLOGY INC.)
Version:    1.11
Speed:  Up to 480Mb/sec
Location ID:    0x14200000 / 8
Current Available (mA): 500
Current Required (mA):  100
Extra Operating Current (mA):   0

    Stereo Vision 2:

    Product ID: 0x9901
    Vendor ID:  0x0ac8  (Z-Star Microelectronics Corporation)
    Version:    27.02
    Serial Number:  SN0099
    Speed:  Up to 480Mb/sec
    Manufacturer:   SHENZHEN RERVISION TECHNOLOGY
    Location ID:    0x14220000 / 10
    Current Available (mA): 500
    Current Required (mA):  500
    Extra Operating Current (mA):   0

    Stereo Vision 2:

    Product ID: 0x9902
    Vendor ID:  0x0ac8  (Z-Star Microelectronics Corporation)
    Version:    27.02
    Serial Number:  SN0100
    Speed:  Up to 480Mb/sec
    Manufacturer:   SHENZHEN RERVISION TECHNOLOGY
    Location ID:    0x14210000 / 9
    Current Available (mA): 500
    Current Required (mA):  500
    Extra Operating Current (mA):   0


  



  But when I list devices, I can see only one [0] :

  



  ffmpeg -f avfoundation -list_devices true -i ""
    [AVFoundation input device @ 0x7fae5b501a80] AVFoundation video devices:
    [AVFoundation input device @ 0x7fae5b501a80] [0] Stereo Vision 2
    [AVFoundation input device @ 0x7fae5b501a80] [1] FaceTime HD Camera
    [AVFoundation input device @ 0x7fae5b501a80] [2] Capture screen 0


  



  capturing from this device captures from one of the cameras.

  



  How can I get ffmpeg to detect the second usb device as well ?

  


• how to make cloud services for camera device iot monitoring and control

  16 novembre 2019, par guardian presence

  Looking to use ffmpeg to pull video trsp stream over tcp pass the video feeds to opencv for object recognition and tracking passing the output to ardunio serial control systems like alarm systems drones ardunio controlled lawn mores.

  Lets say some one has a cctv system and a drone laying around if they connect to my cloud the system can pull the stream from the cctv cameras pull the stream from the drone if object moves in on camera zone the cloud can control the drone to move to that zone and take a closer look at a face if face detected send alert return drone to docking . if ptz camera opencv tracking zooming to face

  if ros can turn into a cloud service need to know about hosting and
  building front end user login and device login.

  i’m a cctv installer new to programming and want to setup cloud login for camera iot devises were users can control and monitoring of devices from a central cloud