Number Plate Detection and Recognition
Vehicles on the road are increasing extensively day by day, particularly in proportion to the industrial revolution and growing economy. The significant use of vehicles has increased the probability of traffic rules violation, causing unexpected accidents, and triggering traffic crimes. Since the po
2025-06-28 16:28:40 - Adil Khan
Number Plate Detection and Recognition
Project Area of Specialization Artificial IntelligenceProject SummaryVehicles on the road are increasing extensively day by day, particularly in proportion to the industrial revolution and growing economy. The significant use of vehicles has increased the probability of traffic rules violation, causing unexpected accidents, and triggering traffic crimes. Since the police officers are unable to check every vehicle on the busy road therefore an automatic intelligent traffic monitoring system is required. The intelligent system can play a vital role in traffic control through the number-plate detection of the vehicles. This research work propose a system for detecting and recognizing of vehicle number plates using a deep learning based convolutional neural network (CNN). The proposed research comprises of two parts: number plate detection and number plate recognition. In the detection part, a vehicle’s image is captured through a digital camera. Then the system segments the number plate region from the image frame. After extracting the number plate region, a super resolution method is applied to convert the low-resolution image into a high-resolution image. The super resolution technique is used with the convolutional layer of CNN to reconstruct the pixel quality of the input image. Each character of the number plate is segmented using a bounding box method. When the number plate detection is done, several features are extracted to represent the number plate and classified using the CNN technique. The novelty of this research is the development of an intelligent system employing CNN to recognize the number-plates (eg: Vehicles), which may have less resolution, and are written in the English language with some specified format. The process of number plate detection can be further distributed into four main steps:
- Preprocessing
- License plate region extraction
- Character Segmentation and Character recognition
Each of the step has its own importance in recognition of number-plate. Image acquired from the camera will be converted into gray scale image and noise removing filters will be used to get the better quality. Finding Exact Location of the License Plate Region is the most important preprocessing step in vehicle detection System because all other steps depends upon the accurate extraction of the region. In the second step, the desired area of number-plate is segmented from the preprocessed image. Segmentation is a process of sub dividing a digital image into its consequent parts. The main purpose to perform segmentation is to find objects and extract meaningful information from digital image. This step is also crucial in number plate detection. The last step of number plate text extraction is character recognition . The text from digital image will be extracted and stored in database.
Project ObjectivesThe Project aims to develop a fast, efficient and effective solution to Number Plate Detection and Recognition (NPDR) of vehicles. It has several applications in our daily life like security system, measure of traffic rules violation, parking system. The detected Number Plate can be used for security of vehicles and to automate parking system. The main objectives of this project are:
- Detection the Number Plate
- Extraction of Text from detected plate
- Enhancement of the Security System of Vehicles
- Automatic Parking at Public and Private Sectors
- Providing a low cost solution
Moreover the research aims to provide customers with Web Application / Mobile Application, Real Time Implementation of system and the organizations using this system will be a source of generating revenue.
Project Implementation MethodThe project solution comprises both the hardware and software. The hardware component includes a high speed computing device that does all the processing of the detected number plate. Besides, a high resolution visual recording device is required to capture the image of vehicle. The captured image is then fed to computing device which will perform processing and will identify vehicle as registered or smuggled. Besides a Web Application will be monitoring all the data for an organization. For parking automation, the project tends to be installed at exit and entry points of Parking Organizations.
Benefits of the ProjectThe benefits include:
- Automatic Parking
- Detection Smuggled vehicles and stolen vehicles
- Identification of unregistered vehicles
- Generate revenue
- An automatic low cost solution
Thefts and Crimes have increased exponentially in some areas of Pakistan. Vehicles are smuggled from different countries and used without paying any tax / custom duty to the government. To cope up with this, the proposed solution tends to be beneficial. The Project uses a registered Vehicles Database API from Safe City Authorities and can identify a vehicle to be smuggled or registered or stolen. Moreover, another application of this project includes Parking Automation. A device will automatically scan the vehicle and save the Park in and Park Out time. In foreign countries, this technology is common everywhere. The project tends to implement a low cost and efficient solution to make it accessible everywhere in Pakistan.
Technical Details of Final DeliverableThe proposed project tends to use a Mobile Application along with a Web application for processing the Number Plate and extracting text from the desired area. The core technology we used in the project is Python. The proposed model uses Google Collab for the training of the Model and testing of the results. Dataset of around 3000 images is used to train the CNN model. It implements Mask-R MobileNet CNN Model (Younis, 2020) which is derived from the TensorFlow API. This Project tends to implement MYSQL as a database for keeping track of the parking records. For Optical Character Recognition (OCR) purposes, a pretrained Model EasyOCR (Kochale, 2021) is available. The project tends to implement OCR using the public API of EasyOCR. The hardware solution involves a fast computing device that can process the data. Besides, a high-quality visual data recording device is required to capture a clear image of the vehicle.
References- (Kochale, 2021) at al. “Real-Time License Plate Detection with EasyOCR”.
2. (Younis, 2020) Younis, Ayesha, et al. "Real-time object detection using pre-trained deep learning models MobileNet-SSD." Proceedings of 2020 the 6th International Conference on Computing and Data Engineering. 2020.
Final Deliverable of the Project HW/SW integrated systemCore Industry ITOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Industry, Innovation and Infrastructure, Partnerships to achieve the GoalRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 80000 | |||
| Computing Device along with GPS visual enabled sensor | Equipment | 1 | 70000 | 70000 |
| Stationary | Miscellaneous | 1 | 10000 | 10000 |