Live Cam Object Identifier

Version: 0.0.1

Receives a live camera URL input and return a screen with the live broadcast image and overlay it with identification for some objects

Dependencies:

Actual version:

{
    "imutils": "0.5.1",
    "numpy": "1.15.4",
    "opencv-python": "3.4.3.18",
    "pypiwin32": "223",
    "pywin32": "224"
}

Future implementations: -- not defined --

Future deprecations:

{
    "pywin32": "224"
}

Live Example:

Not available yet

Simple Local Run:

creates virtual env

pip install -r requiremetns.txt
python src/streamer.py "http://semobrn.sytes.net:1935/live/21.stream/playlist.m3u8"

OBS: this example url is a public security camera from Natal, Brazil

Field	Obrigatory	Description	Input type	Default value
sources	yes	source to have objects identified	Array of Strings	None
queues	no	set queue maximum size for frames	Integer	60
help	no	show this help message and exit	--	--
debug	no	process debug flag	--	--
version	no	output software version	--	--

TODO:

• Get video from an URL and retrieve its frames [OK]
• Implement a thread for video buffering like function [OK]
• Split code on stream, identifier and main in spearated files for better use of python module structure [PENDING]
• Set function for identify array size and fix frame shapes for better visualize multi video input [PENDING]
• Add a multithread to avoid server problems on delay for its video stream [PENDING]
• Identify simple objects like car or cats to verify specific case of use [PENDING]
• Generalize object identification as modules to be passed within command lines or arument on function calls [PENDING]