Speed control on highways and accident prone areas has been a challenging conundrum for government on a global scale. According to reports there are approximately 4 lakh crashes each year causing upto 1 lakh deaths all over the world. A majority of them are caused by overspeeding vehicles which can be controlled by implementing a proper monitoring system that is able to detect overspeeding vehilces.

The idea was to develop a vehicle speed detection system using video streaming. This system requires a video stream which consists of components like vehicles which are in state of motion, starting point and ending point. The system is designed to detect and track the vehicles to calculate the speed of each vehicle. This project is an implementation of Speed Detection System (SDS) which can be used as an alternative for radar and other existing systems. SDS uses several image processing techniques on video stream captured from camera, which makes SDS capable of calculating the speed of moving objects avoiding the traditional radar problems.

SDS process is divided into four phases; first phase is the object detection in which the vehicles that are passing through the road would be detected successfully by using image processing methods. After completion of object detection phase the next step consists of object tracking in which the detected vehicle having assigned ID is tracked over frames by using Euclidean distance formula. Finally from distance and traveled time of detected vehicle, speed of that vehicle is determined by using speed function formula. Lastly, images of each individual vehicle would be captured and stored in desired location. At the end, data visualization and file methods would be used to store and represent all the data that has been collected from object detection, object tracking and speed estimation phase.

There are four main stages of implementation:

• Vehicle detection using image prcessing techniques such as opening, closing and erosion to identify multiple vehicles in the frame.
• Vehicle tracking using euclidean distance formula to track multiple vehicles across the frame and assign them unique IDs.
• Speed estimation using speed formula to identify over-speeding vehicles.
• Data representation for info gathering and visualization.

• OS Windows 10
• Python Programming Language
• Python Libraries (openCv, matplotlib, numpy, dateandtime etc)

• Install latest version of python from here: Download 📥

• Install respective python modules by using following commands:

    pip install opencv-python

    pip install matplotlib

    pip install numpy

• To deploy this project create a SDS folder 📁 inside your E drive.

  

• Create a resources folder 📁 inside SDS folder 📁 and place traffic file 📋 inside. Dwnload traffic.mp4 file from here: Download 📥

  

• Create a TrafficRecord folder 📁 inside SDS folder 📁 and create a exceeded folder 📁 inside it.

  

• This is how your project directory SDS 📁 must look.

  

• Run this command in your terminal

    python main.py

• Vehicle detection and tracking

  

• Storing vehicle images and data

  

• Data report

  

• Graphical representation

  

Speed detection system focuses to detect speed of every vehicle and can monitor as well as reduce the road accidents due to over speeding with proper accuracy and efficiency. As compared to the existing systems speed detection system is easy to handle with less physical effort and is completely automated. Since the cost of this system is many times less than the existing system making it efficicent cost wise. This system also helps to count number of vehicles that pass over the road so that government gets an idea about the maintenance time period of that road. This video-based speed detection system works as good alternative to the existing system. It can also be expanded to various fields of security measures and can be further improved by adding features such as number plate extraction and many more.