Abstract
The recent years have witnessed a considerable rise in the number of vehicles, which has placed transportation infrastructure and traffic control under tremendous pressure. Yielding timely and accurate traffic flow information is essential in the development of traffic control strategies. Despite the continual advances and the wealth of literature available in intelligent transportation system (ITS), there is a lack of practical traffic counting system, which is readily deployable on edge devices. In this study, we introduce a low-cost and effective edge-based system integrating object detection models to perform vehicle detecting, tracking, and counting. First, a vehicle detection dataset (VDD) representing traffic conditions in Vietnam was created. Several deep learning models for VDD were then examined on two different edge device types. Using this detection, we presented a lightweight counting method seamlessly combining with a traditional tracking method to increase counting accuracy. Finally, the traffic flow information is obtained based on counted vehicle categories and their directions. The experiment results clearly indicate that the proposed system achieves the top inference speed at around 26.8 frames per second (FPS) with 92.1% accuracy on the VDD. This proves that our proposal is capable of producing high-accuracy traffic flow information and can be applicable to ITS in order to reduce labor-intensive tasks in traffic management.
Highlights
In the past few years, the intelligent transportation system has become a fundamental branch of smart city construction, playing an essential role in public transport planning, management, and security [1]
Selecting the best vehicle detection model according to both accuracy and speed is likely to be impossible to obtain. is may be explained by the fact that there are twostage detection methods (e.g., R-convolution neural network (CNN), Fast region proposal method with CNNs (R-CNN), Faster R-CNN, and Mask R-CNN) detecting objects through two steps: (1) the model suggests a set of regions of interest based on a region proposal network; (2) the region proposals are dispatched to the channel for bounding-box regression as well as object classification
Dataset Generation. is section presents the vehicle detection dataset (VDD) produced for vehicle detection in real-world conditions. e dataset consists of videos of urban roads and intersection scenes recorded using several traffic cameras. ese videos are minutes in length [64]
Summary
In the past few years, the intelligent transportation system has become a fundamental branch of smart city construction, playing an essential role in public transport planning, management, and security [1]. E framework aims at effectively obtaining traffic flow information via three steps: vehicle detecting, tracking, and counting. (i) We present an AI-based traffic counting framework operating directly on edge devices to obtain traffic flow information effectively. (iv) We show that the proposed framework can obtain high-accuracy traffic flow information in real time while operating on SBCs (acting as edge devices). E regression-based methods aim to exploit global image characteristics (e.g., color histogram and pixel density) to identify and count the number of vehicles. A fuzzy ontology-based lexicon method is exploited to increase the word embedding model’s accuracy Besides these traffic flow variables, many other studies focus on improving information security while transferring data from edge devices to the cloud in ITS. The authors in [49,50,51] exploit the blockchain technology for authentication and authorization, suitable for large-scale intelligent applications
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
