Traffic Load Detection

This is the graduate project of Perach Ben David and Yehuda Aharon Ezer in Software Engineering Master's degree at Ort Braude College.

The project topic is to implement traffic load detection based on YOLO algorithm, The YOLO implementation was took from https://github.com/ultralytics/yolov3.

Installation

Prerequisites

1. Windows 7 / 10 64bit
2. Python 3.7.x 64bit
3. CUDA Toolkit 10.x
4. NVIDIA GPU card (GTX 1050 and above)

Downloading Project

1. Clone this repo:
   git clone https://github.com/aey76/Traffic-Load-Detection.git
2. Download and extract yolo3_v6 and weights from the file Traffic-Load-Detection-yolo3_v6.7z located at 'https://drive.google.com/open?id=1viSKJuym6OeI9MbHxArtm6a1AYzug1s5' into the cloned repo directory.
3. pip install ./yolo3_v6/requirements.txt
4. pip install ./requirements.txt

Running

Run ./gui_main.py

Application Overview

1. Main window layout with view to 3 live webcams

---

2. Set "Identified Vehicles" check box to see the detected vehicles

---

3. Set "Local Loads" check box to see high load areas

---

4. See the green bar raise as the traffic load increase and the traffic load history graph updates

---

5. You can record traffic sessions by pressing the "Start Record" button and reply them later.

---

6. See ./arts/Traffic-Load-Detection_Demo_h265.mp4 video

---