VEHICLE DETECTION, SPEED MONITORING AND CLASSIFICATION FROM VIDEO STREAM
As there is manifold increase in traffic, which demands better road safety measures. One of the main contributing factors in road safety is speed of vehicles which if not observed properly may cause numerous fatal road accidents. Hence there is a dire need to monitor the speed of vehicles to enhance
2025-06-28 16:36:34 - Adil Khan
VEHICLE DETECTION, SPEED MONITORING AND CLASSIFICATION FROM VIDEO STREAM
Project Area of Specialization Artificial IntelligenceProject SummaryAs there is manifold increase in traffic, which demands better road safety measures. One of the main contributing factors in road safety is speed of vehicles which if not observed properly may cause numerous fatal road accidents. Hence there is a dire need to monitor the speed of vehicles to enhance the road safety vis a vis control of accidents. Hence there’s an urgent requirement to devise method through which we can identify the over speeding vehicle. We aim at developing a system
Highways accidents are the leading cause of mortality in teenagers, and the eighth leading cause of death in general globally.
- The installation of HD cameras along with speed Gun on the roads can assist in detection and speed monitoring of vehicle. However, the current system has not been able to capture clear images due to speed mismatch of vehicles and camera. As a result, system is not able to detect its feature i.e. Number Plate of vehicle.
- Focusing on this issue, we propose a system which instead of taking a single picture will record a video stream through which a clear image will be extracted. Furthermore, from that image, features (Number plate, type and color of vehicle) will be extracted using different algorithms which will provide more efficiency and accuracy. Also speed will be recorded using speed Gun.
- VEHICLE NUMBER PLATE DETECTION
High resolution camera makes a short video of a car moving at high speed. From the video a clear frame is selected as a result we will detect a car’s number plate.
- TYPE OF VEHICLE DETECTION
From the shape of car we will detect the type of car i.e. Car, Truck, Bus etc.
- COLOR DETECTION
From the image we will decide the color of car.
- SPEED CHECK
We can monitor the speed of the car.
- GUI (CONTROL PANEL)
We will design Graphical User Interface through which we will control the Speed Gun, Database and other functions.
Project Implementation MethodBy HD camera (Day, Night) the video of vehicle is recorded. By implementing video processing, we will extract a single clear frame on which further image processing is applied for vehicle detection. After detecting a vehicle, we will further detect Number plate of vehicle through which we extract Registration number.
The capture of the images, detection of the vehicle, and the process of recognition are the main steps.
The first step is to take an image of the vehicle at the right time. Thus, the number plate of the vehicle will be visible in the image.
By detecting shape of vehicle, we will identify its type (LTV, HTV). Additionally, we will also identify the colour of the vehicle. For all this implementation we will be using machine learning based video and image processing.
- By using Speed Gun and HD camera we will obtain a stream of frames. Through which a clear frame will be obtained.
- When any vehicle will be in range of speed Gun. Camera will start recording the video for a specific time duration.
- Speed of certain vehicle in range will also be recorded.
- After getting the clear image we will detect the vehicle and further detect the Number plate of the Vehicle. Also we will extract the features of the vehicles i.e. Registration no., Color and type of vehicle. This will be done using YOLO.
- To solve Computer vision problem OpenCV will be used.
- Traffic Control:
Detecting marked blacklisted vehicles, it will be possible to control vehicles that are in search and capture through the installation of the equipment in the main accesses to cities, such as highways and roads.
- Average speed control:
Most speed control devices, such as radars or speed traps, control the vehicle’s instantaneous speed. With this equipment it is possible to control the average speed during an itinerary. By installation of two Number plate recognizing equipment at different points in the same lane, it is possible to make two consecutive recognitions of the number plate and to calculate the average speed of the vehicle.
- Fake Number Plate:
Whenever a faked number plate car is detected, an alert can be generated.
- Traffic optimization:
The installation of Number plate recognizing equipment will allow us to know how many times a vehicle spends to cross an itinerary. This way, the average time can be informed. It is helpful in improving the vehicles mobility during rush hours and traffic jam.
- Vehicle Theft Control:
If any vehicle theft is reported by monitoring the database, we can locate the vehicle easily. It will help the police work more effectively as manual checking system will be automated.
- Car parks:
Recording of registration number will link the number plate and ticket number. Thus, car parking improves their management and security. These are some direct benefits
- Number plate registration: The ticket number, date and time, are linked with number plate in the management application.
- Finding a lost ticket: Thanks to the registry of the number plates it is possible to find a lost ticket and receive the correct amount.
- Interchange is avoided: It is possible to block the exit of a vehicle, if the number plate of the vehicle does not match with the number plate in the entrance ticket.
Data Acquistion Using Hardware
- By using Speed Gun and HD camera we will obtain a stream of frames. Through which a clear frame will be obtained.
- When any vehicle will be in range of speed Gun. Camera will start recording the video for a specific time duration.
- Speed of certain vehicle in range will also be recorded.
Extracting Features Using Deep Learning
- After getting the clear image we will detect the vehicle and further detect the Number plate of the Vehicle. Also we will extract the features of the vehicles i.e. Registration no., Color and type of vehicle. This will be done using YOLO.
- To solve Computer vision problem OpenCV will be used.
Maintaining Database
Following Data will be saved in the Database
- Image of the vehicle
- Registration no.
- Color of vehicle
- Type of vehicle
- Speed
- Time
- Location
We will use MongoDB
Final Deliverable of the Project HW/SW integrated systemCore Industry ITOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Industry, Innovation and InfrastructureRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 53600 | |||
| NVIDIA Jetson Nano | Equipment | 1 | 16000 | 16000 |
| Camera | Equipment | 1 | 30000 | 30000 |
| printing | Miscellaneous | 12 | 550 | 6600 |
| Wires/cables | Equipment | 20 | 50 | 1000 |