People-Detection-and-Counting-In-Out-Line

A people detection using Yolo v3 and Deep Sort to count people who passes a certain line that delimits the entrance of a place. This program can count the people who get into the place or get out.

Project based On

This project is based on the Deep-Sort Algorithm usimg Yolo v3: https://github.com/Qidian213/deep_sort_yolov3 What we are looking forward to resolve in this github is the counting from a line.

Requirements

For CPU:

Keras

tensorflow>=1.4.0 (not compatible with tf 2.0) #This will be resolved soon

NumPy

sklearn

OpenCV

Pillow

---

For GPU:

Keras

tensorflow-gpu (not compatible with tf 2.0) #This will be resolved soon

NumPy

sklearn

OpenCV

Pillow

Download pre-trained weights

Before testing the code you must download the weights for the deep learning model for the trackers and the yoloV3.

Drive Link for Deep Sort Weights : https://drive.google.com/open?id=1rjZpYVorVCDyx25BkkQ0rdr2wdnYM71X

Drive Link for Yolov3 weights: https://drive.google.com/file/d/1uvXFacPnrSMw6ldWTyLLjGLETlEsUvcE/view

You must put both files into the model_data folder

Test

Command Linux Terminal or Anaconda Prompt:

  python main.py -i path_input_video -n 1

Test Video Results: video