/csor-service

Service processing RTSP streams using OpenCV and YOLO Real-Time Object Detection Model.

Primary LanguagePython

CSORS - Camera Stream Object Recognition Service Build Status Python 3.6

This project processes RTSP video streams using OpenCV and the YOLO Real-Time Object Detection Model.

Frames containing a "subject of interest" are posted to a slack channel.

Getting Started

Create Slack Bot Account

Follow Slack documentation for Bot-Users.

Installing Dependencies

src/yolov3-coco/get_model.sh && pip install -r requirements.txt

Running Unit Tests

python -m unittest discover -s test -p "*Test.py"

Usage

Simply pass a suitable RTSP stream URL as an argument when running the application:

python src/main.py -loc http://82.215.168.242:8083/mjpg/video.mjpg

For a full list of arguments run: python src/main.py -h

usage: main.py [-h] [-m MODEL_PATH] [-w WEIGHTS] [-cfg CONFIG] [-i IMAGE_PATH]
               [-v VIDEO_PATH] [-l LABELS] [-c CONFIDENCE] [-th THRESHOLD]
               [-t SHOW_TIME] [-d DEBUG] [-sc SHOW_CAM] [-ri REFRESH_INTERVAL]
               [-loc STREAM_LOCATION] [-stok SLACK_TOKEN]
               [-stit SLACK_ALERT_TITLE] [-schan SLACK_CHANNEL]
               [-tdir TMP_FILE_LOCATION] [-subj SUBJECT_OF_INTEREST]