ank-cha / social-distancing-analyser-covid-19

A social distancing analyzer AI tool to regulate social distancing protocol using video surveillance of CCTV cameras and drones. Social Distancing Analyser to prevent COVID19

Social-Distancing-Analyser-COVID-19

An AI tool to prevent spreading of coronavirus (COVID-19) by using computer vision on video surveillance. A social distancing analyzer AI tool to regulate social distancing protocol using video surveillance of CCTV cameras and drones. Social Distancing Analyser to prevent COVID19

For education purpose only, meant as my contribution towards society

Social Distancing Analyser automatically detects the extent to which social distancing protocols are followed in the area. Deploying it on current surveillance systems and drones used by police to monitor large areas can help to prevent coronavirus by allowing automated and better tracking of activities happening in the area. It shows analytics of the area in real time. It can also be used to alert police in case of considerable violation of social distancing protocols in a particular area.

v1.0 output:

v2.0 output:

V3.0 output (code coming soon)

Features:

• Get the areal time analytics such as:
  • Number of people in a particular area
  • Number of people in high risk
  • The extent of risk to a particular person.
• Doesn't collect any data of a particular person
• Stores a video output for review

Things needed to be improved :

• Auto-calibration [For the given sample video, I've calibrated the model by simulating a 3D depth factor based on the camera position and orientation.] (Check out v2.0)
• Faster processing

Please Note: angle factor is needed to be set between 0 to 1 for v2.0 according to the angle of camera (move towards one as angle becomes verticle)

Installation:

• Fork the repository and download the code.
• Download the following files and place it in the same directory
  • https://github.com/pjreddie/darknet/blob/master/cfg/yolov3.cfg
  • https://pjreddie.com/media/files/yolov3.weights
• For slower CPUs, use yolov3-tiny (link in the code comments)
• Install all the dependenices
• Run social_distancing_analyser.py or social_distancing_analyser 2.0.py

Credits:

@article{yolov3, title={YOLOv3: An Incremental Improvement}, author={Redmon, Joseph and Farhadi, Ali}, journal = {arXiv}, year={2018} }

copyright © 2020 Ankush Chaudhari | All rights reserved

social-distancing-analyser-covid-19's Issues

....

How do you judge whether the social distance is too close？ Can you introduce the calculation formula in detail？

I have a question.How do you judge whether the social distance is too close？
Can you introduce the calculation formula in detail？

How to do the camera calibration

@Ank-Cha

Thanks for your work.

could you please advise how to do the camera calibration in the code?

I tried your code with webcam and close object detection result is poor.

highly appreciate your advice on this.

how can be run it with GPU

Hi Ank thank you for sharing witth us your project , i have i question about GPU:
when i run you code and watch nvidia-smi in an other terminal , the code pythone use juste 41 MiB in the performance of GPU , how can be use all performance , Thanks in advance .

Improvement: Depth Calculation.

Maybe we could join efforts to calculate depth via ROI perspective. Anyone has experience with this?

erorr with default video and code

when i run the code this error display on the screen
Traceback (most recent call last):
File "C:\Users\asus\Desktop\dj\social\social_distancing_analyser.py", line 45, in
ln = [ln[i[0] - 1] for i in net.getUnconnectedOutLayers()]
File "C:\Users\asus\Desktop\dj\social\social_distancing_analyser.py", line 45, in
ln = [ln[i[0] - 1] for i in net.getUnconnectedOutLayers()]
IndexError: invalid index to scalar variable.

Error

[ERROR:0] global /io/opencv/modules/videoio/src/cap.cpp (140) open VIDEOIO(CV_IMAGES): raised OpenCV exception:

OpenCV(4.3.0) /io/opencv/modules/videoio/src/cap_images.cpp:253: error: (-5:Bad argument) CAP_IMAGES: can't find starting number (in the name of file): ./videos/videos in function 'icvExtractPattern'

Processing finished: openop_videos
Traceback (most recent call last):
File "main.py", line 250, in
writer.release()
AttributeError: 'NoneType' object has no attribute 'release'

licencing issue

You are restricting code on Github without an open-source license protecting your work. This means you have little to no power over who can do what with your code without a proper license. You are also publically shaming people who take credit for your code (despite there being a lack of a license, meaning anyone can take credit for your code legally and you can't do a thing) which can be viewed as a violation of Github's terms of use.

I recommend the GNU GPL license (https://www.gnu.org/licenses/gpl-3.0.html) to follow your restrictions, but you may want to review this license to make sure it follows with all of your terms. It is a strong copyleft license that can protect your work.
GNU GPL does allow for commercial use, so consider dual-licensing with the Commons Clause license (https://commonsclause.com/)

Lastly, to fully comply with Github's terms, I recommend not blasting others' social media on Github. You can modify your README to remove the offense, or you can leave it up there and Github can take action against you for DOXXing. It is truly up to you.

Have a fine day and take my advice into consideration!

import cv2

no module named cv2
cant find this module. Again where are dependencies to install please?

how can I use yolov5?

I want to use yolov5 rather than yolov3 , Is there any way?

Doubt

Hey, was going through your code, really cool!
Just wanted to ask how you got the constants in your isclose function for vc_calib_hor, .. c_calib_ver in version2?

angle_factor and H_zoom_factor

Hi Mr. Ank-Cha, I want to ask how you determine the value of "angle_factor" and "H_zoom_factor"

error

in my case it is giving this error ...

(base) C:\Users\Akash Kumar\Downloads\project_ML>D:/anaconda/python.exe "c:/Users/Akash Kumar/Downloads/project_ML/Social-Distancing-Analyser-COVID-19-master/Social-Distancing-Analyser-COVID-19-master/social distancing analyser 2.0.py"
Video name in videos folder: video.mp4
Processing finished: openop_video.mp4
Traceback (most recent call last):
File "c:/Users/Akash Kumar/Downloads/project_ML/Social-Distancing-Analyser-COVID-19-master/Social-Distancing-Analyser-COVID-19-master/social distancing analyser 2.0.py", line 252, in
writer.release()
AttributeError: 'NoneType' object has no attribute 'release'

Video Speed

Hi, My CPU has a NVIDIA geforce 940MX config, still the video lags when I run the code. As suggested I have even tried the same with the tiny versions, but the speed improved by 10-20 % only. Can you suggest if anywhere else it has to be tuned in your code to obtain the correct fps.

Thanks