VIDEO SURVEILLANCE-BASED TRAFFIC MANAGEMENT SYSTEM
This significant increase enforced modern transportation system to promote the performance of tra?c controlling system. Therefore, in order to maintain the tra?c e?ectively and safely, automation and arti?cial intelligence have become the mainstream. In this regard we are developing a system that is
2025-06-28 16:36:36 - Adil Khan
VIDEO SURVEILLANCE-BASED TRAFFIC MANAGEMENT SYSTEM
Project Area of Specialization Artificial IntelligenceProject SummaryThis significant increase enforced modern transportation system to promote the performance of tra?c controlling system. Therefore, in order to maintain the tra?c e?ectively and safely, automation and arti?cial intelligence have become the mainstream. In this regard we are developing a system that is able to solve traffic problems such as the speed estimation, car direction, traffic density, traffic rule violation etc. In future, the system also perform automatic detection of bike-riders with or without helmet using surveillance videos in real time.
Project ObjectivesGeneral Capabilities of Project:
- Detection and classification of the vehicles
- Vehicle Counting
- Vehicle color prediction
- Vehicle speed estimation
- License Plate Character Recognition
- Drone view detection
- Traffic Collision
Firstly we take different videos from those videos we classified object of interest using supervised learning fashion and then we applied machine learning algorithm , the module of vehicle detection based on machine learning is extremely effective for roads with heavy traffic flow after it we train detector for each class(bike, car, rickshaw, etc.) by "Faster RCNN model” which is developed on TensorFlow. The idea which we proposed processes an input video to track and detects the vehicle through its motion, classification, the speed, traffic density, traffic rules violation and also counts the total number of vehicles on the road. Number plate identification is also being incorporated for identifying certain cars and the distance between driving car and others is also calculated.
To enhance the process, we use consolidation of different image processing and computer vision techniques.The TensorFlow Object Counting API is used as a base for object counting on this project. Tensor Flow’s Object Detection API is an open source framework built on top of Tensor Flow that makes it easy to construct, train and deploy object detection models. This is a loop that continue working till reaching end of the video. The main pipeline of the tracker is given in the Figure below.
The traffic situations have undergone dramatic changes over the last few years and this leads to the development of an innovative technology to detect and count vehicles or other vulnerable road users. This project is focused on designing a method for vehicle detection which in turn can reduce the traffic jam. Even though, many projects were being carried out for vehicle detection, most of the methods had drawbacks or false detection (FP).
Additionally, the innovative technology will be applied in other systems, such as traffic surveillance, toll collection and parking lot access control.
Technical Details of Final DeliverableSotware Details
- TensorFlow is an open source software library for numerical computation using data flow graphs. Nodes in the graph represent mathematical operations, while the graph edges represent the multidimensional data arrays (tensors) communicated between them.
- OpenCV (Open Source Computer Vision Library is an open source computer vision and machine learning software library. OpenCV was built to provide a common infrastructure for computer vision applications and to accelerate the use of machine perception in the commercial products.
- FASTER R-CNN Our object detection system, called Faster R-CNN, is composed of two modules. The ?rst module is a deep fully convolutional network that proposes regions, and the second module is the Fast R-CNN detector that uses the proposed regions.
Creating accurate Machine Learning Models which are capable of identifying and localizing multiple objects in a single image remained a core challenge in computer vision. But, with recent advancements in Deep Learning, Object Detection applications are easier to develop than ever before. Tensor Flow’s Object Detection API is an open source framework built on top of Tensor Flow that makes it easy to construct, train and deploy object detection models.
Our project is divided into three major components:
- Vehicle Detection and Counting,
- License Plate Character Recognition.
Phase I is detecting whether an input contains vehicles or not. If the algorithm predicts that the input contains vehicles, then we need to define a method which will precisely locate and crop the vehicle from the original data. Starting with the output from Phase I as the input, Phase II should search for the vehicle’s license plate and have the ability to recognize the number on the car plate.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 70000 | |||
| Dahua and hikvision HD CCTV System | Equipment | 1 | 15000 | 15000 |
| GIGABYTE GeForce RTX 2060 DirectX 12 6GB 192-Bit GDDR6 Graphics Card | Equipment | 1 | 55000 | 55000 |