Merge pull request datitran#44 from samcrane8/master

HLS Stream Object Detection

Author: datitran

README.md

@@ -4,13 +4,15 @@ A real-time object recognition application using [Google's TensorFlow Object Det
 
 ## Getting Started
 1. `conda env create -f environment.yml`
-2. `python object_detection_app.py`
+2. `python object_detection_app.py` / `python object_detection_multithreading.py`
  Optional arguments (default value):
  * Device index of the camera `--source=0`
  * Width of the frames in the video stream `--width=480`
  * Height of the frames in the video stream `--height=360`
  * Number of workers `--num-workers=2`
  * Size of the queue `--queue-size=5`
+ * Get video from HLS stream rather than webcam '--stream-input=http://somertmpserver.com/hls/live.m3u8'
+ * Send stream to livestreaming server '--stream-output=--stream=http://somertmpserver.com/hls/live.m3u8'
 
 ## Tests
 ```

environment-linux.yaml

@@ -0,0 +1,48 @@
+name: object-detection
+channels: !!python/tuple
+- menpo
+- defaults
+dependencies:
+- cycler=0.10.0=py35_0
+- freetype=2.5.5=2
+- icu=54.1=0
+- jbig=2.1=0
+- menpo::opencv3=3.0.1=py35_0
+- jpeg=9b=0
+- libpng=1.6.27=0
+- libtiff=4.0.6=3
+- matplotlib=2.0.2=np113py35_0
+- menpo::tbb
+- mkl=2017.0.1=0
+- numpy=1.13.0=py35_0
+- olefile=0.44=py35_0
+- openssl=1.0.2l=0
+- pillow=4.2.1=py35_0
+- pip=9.0.1=py35_1
+- py=1.4.34=py35_0
+- pyparsing=2.2.0=py35_0
+- pyqt=5.6.0=py35_2
+- pytest=3.2.1=py35_0
+- python=3.5.3=1
+- python-dateutil=2.6.1=py35_0
+- pytz=2017.2=py35_0
+- qt=5.6.2
+- readline=6.2=2
+- setuptools=27.2.0=py35_0
+- sip=4.18=py35_0
+- six=1.10.0=py35_0
+- sqlite=3.13.0=0
+- tk=8.5.18=0
+- wheel=0.29.0=py35_0
+- xz=5.2.2=1
+- zlib=1.2.8=3
+- pip:
+ - backports.weakref==1.0rc1
+ - bleach==1.5.0
+ - html5lib==0.9999999
+ - markdown==2.2.0
+ - protobuf==3.3.0
+ - tensorflow==1.2.0
+ - werkzeug==0.12.2
+prefix: /Users/datitran/anaconda/envs/object-detection
+

environment.yml

@@ -44,5 +44,6 @@ dependencies:
  - protobuf==3.3.0
  - tensorflow==1.2.0
  - werkzeug==0.12.2
+ - scipy
 prefix: /Users/datitran/anaconda/envs/object-detection
 

object_detection_app.py

@@ -6,7 +6,7 @@
 import numpy as np
 import tensorflow as tf
 
-from utils.app_utils import FPS, WebcamVideoStream
+from utils.app_utils import FPS, WebcamVideoStream, HLSVideoStream
 from multiprocessing import Queue, Pool
 from object_detection.utils import label_map_util
 from object_detection.utils import visualization_utils as vis_util
@@ -85,6 +85,7 @@ def worker(input_q, output_q):
 
 if __name__ == '__main__':
  parser = argparse.ArgumentParser()
+ parser.add_argument('-str', '--stream', dest="stream", action='store', type=str, default=None)
  parser.add_argument('-src', '--source', dest='video_source', type=int,
  default=0, help='Device index of the camera.')
  parser.add_argument('-wd', '--width', dest='width', type=int,
@@ -104,9 +105,17 @@ def worker(input_q, output_q):
  output_q = Queue(maxsize=args.queue_size)
  pool = Pool(args.num_workers, worker, (input_q, output_q))
 
- video_capture = WebcamVideoStream(src=args.video_source,
+
+ if (args.stream):
+ print('Reading from hls stream.')
+ video_capture = HLSVideoStream(src=args.stream).start()
+ else:
+ print('Reading from webcam.')
+ video_capture = WebcamVideoStream(src=args.video_source,
  width=args.width,
  height=args.height).start()
+
+ 
  fps = FPS().start()
 
  while True: # fps._numFrames < 120

object_detection_multithreading.py

@@ -3,11 +3,13 @@
 import time
 import argparse
 import numpy as np
+import subprocess as sp
+import json
 import tensorflow as tf
 
 from queue import Queue
 from threading import Thread
-from utils.app_utils import FPS, WebcamVideoStream, draw_boxes_and_labels
+from utils.app_utils import FPS, HLSVideoStream, WebcamVideoStream, draw_boxes_and_labels
 from object_detection.utils import label_map_util
 
 CWD_PATH = os.getcwd()
@@ -83,22 +85,29 @@ def worker(input_q, output_q):
 
 if __name__ == '__main__':
  parser = argparse.ArgumentParser()
+ parser.add_argument('-strin', '--stream-input', dest="stream_in", action='store', type=str, default=None)
  parser.add_argument('-src', '--source', dest='video_source', type=int,
  default=0, help='Device index of the camera.')
  parser.add_argument('-wd', '--width', dest='width', type=int,
- default=480, help='Width of the frames in the video stream.')
+ default=640, help='Width of the frames in the video stream.')
  parser.add_argument('-ht', '--height', dest='height', type=int,
- default=360, help='Height of the frames in the video stream.')
+ default=480, help='Height of the frames in the video stream.')
+ parser.add_argument('-strout','--stream-output', dest="stream_out", help='The URL to send the livestreamed object detection to.')
  args = parser.parse_args()
 
- input_q = Queue(5) # fps is better if queue is higher but then more lags
+ input_q = Queue(1) # fps is better if queue is higher but then more lags
  output_q = Queue()
  for i in range(1):
  t = Thread(target=worker, args=(input_q, output_q))
  t.daemon = True
  t.start()
 
- video_capture = WebcamVideoStream(src=args.video_source,
+ if (args.stream_in):
+ print('Reading from hls stream.')
+ video_capture = HLSVideoStream(src=args.stream_in).start()
+ else:
+ print('Reading from webcam.')
+ video_capture = WebcamVideoStream(src=args.video_source,
  width=args.width,
  height=args.height).start()
  fps = FPS().start()
@@ -125,7 +134,10 @@ def worker(input_q, output_q):
  int(point['ymin'] * args.height) - 10), color, -1, cv2.LINE_AA)
  cv2.putText(frame, name[0], (int(point['xmin'] * args.width), int(point['ymin'] * args.height)), font,
  0.3, (0, 0, 0), 1)
- cv2.imshow('Video', frame)
+ if args.stream_out:
+ print('Streaming elsewhere!')
+ else:
+ cv2.imshow('Video', frame)
 
  fps.update()