/human_detection

Primary LanguagePython

Human Detection examples

A set of simple Python scripts for accessing camera video feed (pc camera / rtsp video feed) via OpenCV & simple Human Detection scenarios.

The best script is the Tensorflow based example in the tf_example folder.

Libraries used:

  • flask (for web based app)
  • opencv-python (pip)
  • tensorflow
  • numpy

Dockerized Human Detection service with web access

Build the service

After cloning the project, go to the docker directory and run:

$ docker build -t human_detection .

in order to build the service.

Exposed ports: 5000 (for web access).

For building the same dockerized service with enabled Gstreamer backend support, run:

$ docker build -t human_detection_gst -f ./Dockerfile-gst .

Exposed ports: 5001 (for web access), 8002 (for receiving UDP video stream).

Run the service

Start the container with this command:

$ docker run --rm -it --net=host --entrypoint "/bin/bash" human_detection

or:

$ docker run --rm -it --net=host --entrypoint "/bin/bash" human_detection_gst

depending on the service you wish to run.

Inside the container, navigate to the ./human_detection/tf_example directory, edit the RCNN_inception_v2_human_detection_rtsp_web.py with the correct RTSP link and, lastly run:

$ python3 RCNN_inception_v2_human_detection_rtsp_web.py

The same scripts can be used for obtaining video feed with Gstreamer backend, by editing the cap = cv2.VideoCapture() line with the correct Gstreamer pipeline. A sample pipeline can be found in the pc_camera_web.py script.

You can then access the web service via this URL: http://localhost:5000 or http://localhost:5000/video_feed .