Awesome Open Source
Awesome Open Source


People Counting in Real-Time using live video stream/IP camera in OpenCV.

This is an improvement/modification to

Refer to added Features. Also, added support for an IP camera.

Live demo

  • The primary aim is to use the project as a business perspective, ready to scale.
  • Use case: counting the number of people in the stores/buildings/shopping malls etc., in real-time.
  • Sending an alert to the staff if the people are way over the limit.
  • Automating features and optimising the real-time stream for better performance (with threading).
  • Acts as a measure towards footfall analysis and in a way to tackle COVID-19.

Table of Contents

Simple Theory

SSD detector:

  • We are using a SSD (Single Shot Detector) with a MobileNet architecture. In general, it only takes a single shot to detect whatever is in an image. That is, one for generating region proposals, one for detecting the object of each proposal.
  • Compared to other 2 shot detectors like R-CNN, SSD is quite fast.
  • MobileNet, as the name implies, is a DNN designed to run on resource constrained devices. For example, mobiles, ip cameras, scanners etc.
  • Thus, SSD seasoned with a MobileNet should theoretically result in a faster, more efficient object detector.

Centroid tracker:

  • Centroid tracker is one of the most reliable trackers out there.
  • To be straightforward, the centroid tracker computes the centroid of the bounding boxes.
  • That is, the bounding boxes are (x, y) co-ordinates of the objects in an image.
  • Once the co-ordinates are obtained by our SSD, the tracker computes the centroid (center) of the box. In other words, the center of an object.
  • Then an unique ID is assigned to every particular object deteced, for tracking over the sequence of frames.

Running Inference

  • First up, install all the required Python dependencies:
pip install -r requirements.txt

The requirements will be updated timely, but note that there can always be version conflicts between the dependencies themselves and other factors like OS, hardware etc.

  • To run inference on a test video file, head into the directory/use the command:
python --prototxt mobilenet_ssd/MobileNetSSD_deploy.prototxt --model mobilenet_ssd/MobileNetSSD_deploy.caffemodel --input videos/example_01.mp4
  • To run inference on an IP camera, first setup your camera url in 'mylib/':
# Enter the ip camera url (e.g., url = '')
url = ''
  • Then run with the command:
python --prototxt mobilenet_ssd/MobileNetSSD_deploy.prototxt --model mobilenet_ssd/MobileNetSSD_deploy.caffemodel

Set url = 0 for webcam.


The following is an example of the added features. Note: You can easily on/off them in the config. options (mylib/

1. Real-Time alert:

  • If selected, we send an email alert in real-time. Use case: If the total number of people (say 10 or 30) exceeded in a store/building, we simply alert the staff.
  • You can set the max. people limit in config. (Threshold = 10).
  • This is pretty useful considering the COVID-19 scenario.
  • Note: To setup the sender email, please refer the instructions inside 'mylib/'. Setup receiver email in the config.

2. Threading:

  • Multi-Threading is implemented in 'mylib/'. If you ever see a lag/delay in your real-time stream, consider using it.
  • Threading removes OpenCV's internal buffer (which basically stores the new frames yet to be processed until your system processes the old frames) and thus reduces the lag/increases fps.
  • If your system is not capable of simultaneously processing and outputting the result, you might see a delay in the stream. This is where threading comes into action.
  • It is most suitable for solid performance on complex real-time applications. To use threading:

set Thread = True in config.

3. Scheduler:

  • Automatic scheduler to start the software. Configure to run at every second, minute, day, or Monday to Friday.
  • This is extremely useful in a business scenario, for instance, you can run it only at your desired time (9-5?).
  • Variables and memory would be reset == less load on your machine.
##Runs at every day (9:00 am). You can change it.

4. Timer:

  • Configure stopping the software after a certain time, e.g., 30 min or 9 hours from now.
  • All you have to do is set your desired time and run the script.
if Timer:
	# Automatic timer to stop the live stream. Set to 8 hours (28800s).
	t1 = time.time()
	if num_seconds > 28800:

5. Simple log:

  • Logs all data at end of the day.
  • Useful for footfall analysis.




Next steps

  • Train the SSD on human data (with a top-down view).
  • Experiment with other detectors and benchmark the results on computationally less expensive embedded hardware.
  • Evaluate the performance on multiple IP cameras.


Thanks for the read & have fun!

To get started/contribute quickly (optional) ...

  • Option 1

    • 🍴 Fork this repo and pull request!
  • Option 2

    • 👯 Clone this repo:
    $ git clone
  • Roll it!

saimj7/ 19-08-2020 © Sai_Mj.

Alternative Project Comparisons
Related Awesome Lists
Top Programming Languages

Get A Weekly Email With Trending Projects For These Topics
No Spam. Unsubscribe easily at any time.
Python (889,550
Deep Learning (39,305
Camera (16,798
Opencv (13,690
Real Time (10,690
Covid 19 (4,632
Ssd (1,973
Threading (1,481
Mobilenet (1,161
Ip Camera (129
People Counter (31
Centroid Tracking (4