Nikhita Tithi, Harika Chadalavada, Shubham Patil, Gagneet Sachdeva
Demonstration Video - https://www.youtube.com/watch?v=vEZsc6IHTiQ
Security surveillance using vehicle license plate detection.
The project's objective is to create an intelligent security system that increases the effectiveness of security by detecting license plates using OpenCV. The system will process the frames from a video stream as input to find passing automobiles' license plates. To identify and handle problems like illegal parking or toll collection, which can increase congestion and decrease traffic efficiency, the detected license plate numbers will be used.
The system will be built to process video streams from cameras placed at critical points along roads and highways in real time. To find license plates, computer vision techniques will be applied to the video streams. To find any unauthorized vehicles that are parked in prohibited locations and give a prompt to the system by adding that number to the database.
The system will be educated using advanced computer vision methods to assure accuracy and dependability. As a result, as it analyzes more data, the system will get better over time, becoming more precise and dependable. The system will be able to respond to problems like unauthorized parking or toll deduction more quickly and effectively. The system's advantages include increased safety, lessened congestion, and increased transportation effectiveness. By identifying unregistered vehicles and stopping them from entering forbidden areas or causing accidents, the technology can also help to increase safety.
The project will entail creating computer vision algorithms for system training and improvement, computer vision algorithms for license plate detection, integrating the system with current transportation infrastructure, and testing the system in real-world scenarios.
The system will take video streams as input from cameras placed at critical points along roads and highways. These cameras will capture passing vehicles, and the system will use computer vision techniques to detect license plates from the video frames. As different regions have different license plate formats, we will feed the system with the license plate format for the region it is deployed to alarm for false license plates.
The system will output the license plate numbers of passing vehicles in real time. This information can be used to identify vehicles, track them, and take action if necessary. The system will prompt an alert if it detects any unauthorized vehicles parked in prohibited locations. This prompt can be sent to the authorities responsible for enforcing parking regulations. We can also integrate this system with the current transportation infrastructure for handling toll collection and parking fee payment. It can automatically deduct the toll or parking fee from the driver's account, thereby reducing congestion at toll plazas and parking lots. The system can also be used to send real-time alerts to law enforcement agencies in case of any suspicious activities or wanted vehicles. This information can be used to take immediate action to prevent any potential threats.
Gathering Ground Truth Data: Gathering a set of ground truth data, such as pictures or videos of automobiles with license plates, would be the initial stage in assessing the software program. This information can be gathered from a variety of places, including online databases and CCTV cameras.
The ground truth data can then be divided into training and testing datasets to prepare the test data. Using OpenCV and computer vision, the license plate detection algorithm may be trained using the training dataset. The algorithm's performance can then be assessed using the testing dataset.
Measuring Performance: A number of metrics, including precision, recall, and F1 score, can be used to measure the effectiveness of the license plate detecting algorithm. These scores give a broad indication of how accurately the system can identify license plates.
Conducting Experiments: Experiments can be carried out to examine how well the algorithm performs in a variety of settings, including various illumination, camera angles, and weather conditions. These tests will aid in identifying the algorithm's shortcomings and potential areas for development.
Algorithm Refinement: The license plate detecting algorithm can be enhanced to perform better based on the analysis of the results. Changing certain settings, like the threshold for identifying license plates, or including fresh features, such as machine learning techniques, may be necessary to gradually increase accuracy.
Repeating the Evaluation: To assess the effectiveness of the improved algorithm, the evaluation procedure can be carried out again. The method can be tested, examined, and improved in this iterative manner until the required level of performance and accuracy is attained.
The given code executes a number of image processing operations, such as contrast amplification, thresholding, contour detection, and masking. On the other hand, prospective areas of interest (ROI) that might include the number plate can be extracted using the contour detection process.
Research and finalize project idea and scope Skimming Research papers suggested in Canvas Review of existing research related to Object detection and on our topic “Security surveillance using vehicle license plate detection.” Create a Project Proposal
Detailed Literature review Setting up a development environment and beginning preprocessing the data
Submission of Project Design Document on May 15
finalize methods and techniques needed to achieve our problem statement, outline expected outputs
Preprocessing, training for the model, and Implementation of the object detection algorithm
Get significant improvements in code
Prepare an outline for the design document and get it ready.
Implement various OCR techniques, Integration, and develop analytic components and ensure the readiness of the code. Write Final Report
Project Demonstration