scholarly journals Vehicle Detection with Self-Training for Adaptative Video Processing Embedded Platform

2020 ◽  
Vol 10 (17) ◽  
pp. 5763
Author(s):  
Sebastian Cygert ◽  
Andrzej Czyżewski
Keyword(s):  
Video Processing ◽  
Vehicle Detection ◽  
Traffic Monitoring ◽  
Fine Tuning ◽  
Estimation Accuracy ◽  
Real Time Processing ◽  
Embedded Platform ◽  
Human Effort ◽  
The Subject ◽  
Sort Algorithm

Traffic monitoring from closed-circuit television (CCTV) cameras on embedded systems is the subject of the performed experiments. Solving this problem encounters difficulties related to the hardware limitations, and possible camera placement in various positions which affects the system performance. To satisfy the hardware requirements, vehicle detection is performed using a lightweight Convolutional Neural Network (CNN), named SqueezeDet, while, for tracking, the Simple Online and Realtime Tracking (SORT) algorithm is applied, allowing for real-time processing on an NVIDIA Jetson Tx2. To allow for adaptation of the system to the deployment environment, a procedure was implemented leading to generating labels in an unsupervised manner with the help of background modelling and the tracking algorithm. The acquired labels are further used for fine-tuning the model, resulting in a meaningful increase in the traffic estimation accuracy, and moreover, adding only minimal human effort to the process allows for further accuracy improvement. The proposed methods, and the results of experiments organised under real-world test conditions are presented in the paper.

scholarly journals FGSC: Fuzzy Guided Scale Choice SSD Model for Edge AI Design on Real-Time Vehicle Detection and Class Counting

Sensors ◽  
2021 ◽  
Vol 21 (21) ◽  
pp. 7399
Author(s):  
Ming-Hwa Sheu ◽  
S M Salahuddin Morsalin ◽  
Jia-Xiang Zheng ◽  
Shih-Chang Hsia ◽  
Cheng-Jian Lin ◽  
...  
Keyword(s):  
Neural Network ◽  
Real Time ◽  
Network Architecture ◽  
Deep Neural Network ◽  
Vehicle Detection ◽  
Traffic Monitoring ◽  
Sigmoid Function ◽  
Real Time Processing ◽  
Scale Choice ◽  
Training Stage

The aim of this paper is to distinguish the vehicle detection and count the class number in each classification from the inputs. We proposed the use of Fuzzy Guided Scale Choice (FGSC)-based SSD deep neural network architecture for vehicle detection and class counting with parameter optimization. The 'FGSC' blocks are integrated into the convolutional layers of the model, which emphasize essential features while ignoring less important ones that are not significant for the operation. We created the passing detection lines and class counting windows and connected them with the proposed FGSC-SSD deep neural network model. The 'FGSC' blocks in the convolution layer emphasize essential features and find out unnecessary features by using the scale choice method at the training stage and eliminate that significant speedup of the model. In addition, FGSC blocks avoided many unusable parameters in the saturation interval and improved the performance efficiency. In addition, the Fuzzy Sigmoid Function (FSF) increases the activation interval through fuzzy logic. While performing operations, the FGSC-SSD model reduces the computational complexity of convolutional layers and their parameters. As a result, the model tested Frames Per Second (FPS) on edge artificial intelligence (AI) and reached a real-time processing speed of 38.4 and an accuracy rate of more than 94%. Therefore, this work might be considered an improvement to the traffic monitoring approach by using edge AI applications.

scholarly journals Automatic Vehicles Detection, Classification and Counting Techniques / Survey

2020 ◽  
pp. 1811-1822
Author(s):  
Mustafa Najm ◽  
Yossra Hussein Ali
Keyword(s):  
Real Time ◽  
Intelligent Transportation Systems ◽  
Background Subtraction ◽  
Moving Objects ◽  
Vehicle Detection ◽  
Intelligent Transportation ◽  
Traffic Monitoring ◽  
Transportation Systems ◽  
Traffic Condition ◽  
Real Time Processing

Vehicle detection (VD) plays a very essential role in Intelligent Transportation Systems (ITS) that have been intensively studied within the past years. The need for intelligent facilities expanded because the total number of vehicles is increasing rapidly in urban zones. Traffic monitoring is an important element in the intelligent transportation system, which involves the detection, classification, tracking, and counting of vehicles. One of the key advantages of traffic video detection is that it provides traffic supervisors with the means to decrease congestion and improve highway planning. Vehicle detection in videos combines image processing in real-time with computerized pattern recognition in flexible stages. The real-time processing is very critical to keep the appropriate functionality of automated or continuously working systems. VD in road traffics has numerous applications in the transportation engineering field. In this review, different automated VD systems have been surveyed,  with a focus on systems where the rectilinear stationary camera is positioned above intersections in the road rather than being mounted on the vehicle. Generally, three steps are utilized to acquire traffic condition information, including background subtraction (BS), vehicle detection and vehicle counting. First, we illustrate the concept of vehicle detection and discuss background subtraction for acquiring only moving objects. Then a variety of algorithms and techniques developed to detect vehicles are discussed beside illustrating their advantages and limitations. Finally, some limitations shared between the systems are demonstrated, such as the definition of ROI, focusing on only one aspect of detection, and the variation of accuracy with quality of videos. At the point when one can detect and classify vehicles, then it is probable to more improve the flow of the traffic and even give enormous information that can be valuable for many applications in the future.

scholarly journals Orientation- and Scale-Invariant Multi-Vehicle Detection and Tracking from Unmanned Aerial Videos

2019 ◽  
Vol 11 (18) ◽  
pp. 2155 ◽  
Author(s):  
Jie Wang ◽  
Sandra Simeonova ◽  
Mozhdeh Shahbazi
Keyword(s):  
Object Tracking ◽  
Traffic Flow ◽  
State Of The Art ◽  
Vehicle Detection ◽  
Traffic Monitoring ◽  
Fine Tuning ◽  
Processing Technologies ◽  
Detection And Tracking ◽  
Vehicle Detection And Tracking ◽  
Tracking By Detection

Along with the advancement of light-weight sensing and processing technologies, unmanned aerial vehicles (UAVs) have recently become popular platforms for intelligent traffic monitoring and control. UAV-mounted cameras can capture traffic-flow videos from various perspectives providing a comprehensive insight into road conditions. To analyze the traffic flow from remotely captured videos, a reliable and accurate vehicle detection-and-tracking approach is required. In this paper, we propose a deep-learning framework for vehicle detection and tracking from UAV videos for monitoring traffic flow in complex road structures. This approach is designed to be invariant to significant orientation and scale variations in the videos. The detection procedure is performed by fine-tuning a state-of-the-art object detector, You Only Look Once (YOLOv3), using several custom-labeled traffic datasets. Vehicle tracking is conducted following a tracking-by-detection paradigm, where deep appearance features are used for vehicle re-identification, and Kalman filtering is used for motion estimation. The proposed methodology is tested on a variety of real videos collected by UAVs under various conditions, e.g., in late afternoons with long vehicle shadows, in dawn with vehicles lights being on, over roundabouts and interchange roads where vehicle directions change considerably, and from various viewpoints where vehicles’ appearance undergo substantial perspective distortions. The proposed tracking-by-detection approach performs efficiently at 11 frames per second on color videos of 2720p resolution. Experiments demonstrated that high detection accuracy could be achieved with an average F1-score of 92.1%. Besides, the tracking technique performs accurately, with an average multiple-object tracking accuracy (MOTA) of 81.3%. The proposed approach also addressed the shortcomings of the state-of-the-art in multi-object tracking regarding frequent identity switching, resulting in a total of only one identity switch over every 305 tracked vehicles.

Unmanned Aerial Vehicle–Based Traffic Analysis: Methodological Framework for Automated Multivehicle Trajectory Extraction

2017 ◽  
Vol 2626 (1) ◽  
pp. 25-33 ◽  
Author(s):  
Muhammad Arsalan Khan ◽  
Wim Ectors ◽  
Tom Bellemans ◽  
Davy Janssens ◽  
Geert Wets
Keyword(s):  
Traffic Safety ◽  
Video Processing ◽  
Traffic Monitoring ◽  
Future Research ◽  
Methodological Framework ◽  
Vehicle Detection And Tracking ◽  
Potential Applications ◽  
Multiple Vehicles ◽  
Processing Techniques ◽  
Optimal Data Analysis

Unmanned aerial vehicles (UAVs), commonly referred to as drones, are one of the most dynamic and multidimensional emerging technologies of the modern era. This technology has recently found multiple potential applications within the transportation field, ranging from traffic surveillance applications to traffic network analysis. To conduct a UAV-based traffic study, extremely diligent planning and execution are required followed by an optimal data analysis and interpretation procedure. In this study, however, the main focus was on the processing and analysis of UAV-acquired traffic footage. A detailed methodological framework for automated UAV video processing is proposed to extract the trajectories of multiple vehicles at a particular road segment. Such trajectories can be used either to extract various traffic parameters or to analyze traffic safety situations. The proposed framework, which provides comprehensive guidelines for an efficient processing and analysis of a UAV-based traffic study, comprises five components: preprocessing, stabilization, georegistration, vehicle detection and tracking, and trajectory management. Until recently, most traffic-focused UAV studies have employed either manual or semiautomatic processing techniques. In contrast, this paper presents an in-depth description of the proposed automated framework followed by a description of a field experiment conducted in the city of Sint-Truiden, Belgium. Future research will mainly focus on the extension of the applications of the proposed framework in the context of UAV-based traffic monitoring and analysis.

The Gathering of A Community: The British-born of San Francisco in 1852 and 1872

1976 ◽  
Vol 10 (3) ◽  
pp. 279-312 ◽  
Author(s):  
R. A. Burchell
Keyword(s):  
Nineteenth Century ◽  
San Francisco ◽  
High Rate ◽  
Geographical Mobility ◽  
The Third ◽  
Human Effort ◽  
The Subject ◽  
Incomplete Picture ◽  
Do So

Studies of the Massachusetts communities of Newburyport and Boston have revealed a high rate of geographical mobility for their populations, in excess of what had been previously thought. Because of the difficulty in tracing out-migrants these works have concentrated on persisters, though to do so is to give an incomplete picture of communal progress. Peter R. Knights in his study of Boston between 1830 and 1860 attempted to follow his out-migrants but was only able to trace some 27 per cent of them. The problem of out-migration is generally regarded as being too large for solution through human effort, but important enough now to engage the computer. What follows bears on the subject of out-migration, for it is an analysis of where part of the migrating populations of the east went in the third quarter of the nineteenth century, namely to San Francisco.

scholarly journals Hakikat Ilmu Pengetahuan dalam Perspektif Modern dan Islam

2017 ◽  
Vol 4 (2) ◽  
pp. 202-217
Author(s):  
Siti Makhmudah
Keyword(s):  
Human Experience ◽  
Coherence Theory ◽  
Theory Theory ◽  
Human Thought ◽  
Sources Of Knowledge ◽  
Human Effort ◽  
The Subject ◽  
Understanding Of Science ◽  
Structure Of Knowledge ◽  
Human Source

Science in a unity appeared in dimensional. Philosophy is an activity though human thought guided their efforts on finding cause for over everything and how human effort after learning of the matter. This research aims to understand: (1) knowledge and understanding of science in etymology and terminology; (2) perbedan of science, knowledge and religion in epistimologi; (3) the extent to which science in Islam; (4) the principal traits of science; (5) the theory of truth; (6) the sources of knowledge; (7) the boundaries of science; (8) the structure of knowledge. Results of the study can be described in several options, which are: first, science is the summary of a set of knowledge or the result of knowledge and facts. While religion is a belief or faith tata tata over something that is absolutely beyond human, appropriate and in line with the faith and worship. Second, with regard to the characteristics of the subject matter of science is as follows: 1) Systematically; 2) Generality; 3) Rationality; 4) Objectivity; 5) Verifiabilitas, 6) and Communality. Third, in Theory a theory of truth is no 3: the theory of correspondence, coherence Theory, theory of pragmatism. Fourth, human source of knowledge using two ways to obtain the correct knowledge, first through ratio and secondly through experience. Fifth, limiting his explorations in the science of human experience, thus embarking upon science exploration on human experience and stop on the human experience, and that is the limits of science. Sixth, the science is essentially a collection of knowledge that is explaining the various symptoms of nature which allows a human doing a series of actions to control these symptoms based on the explanation there is.

Distributed Video Coding and Content Analysis for Resource Constraint Multimedia Applications

2012 ◽  
pp. 251-273
Author(s):  
Praveen Kumar ◽  
Amit Pande ◽  
Ankush Mittal ◽  
Abhisek Mudgal
Keyword(s):  
Video Coding ◽  
Real Time ◽  
Video Processing ◽  
Mobile Agents ◽  
Video Transmission ◽  
Moving Objects ◽  
Individual Case ◽  
Multimedia Applications ◽  
Real Time Processing ◽  
Laboratory Setup

Video coding and analysis for low power and low bandwidth multimedia applications has always been a great challenge. The limited computational resources on ubiquitous multimedia devices like cameras along with low and varying bandwidth over wireless network lead to serious bottlenecks in delivering real-time streaming of videos for such applications. This work presents a Content-based Network-adaptive Video-transmission (CbNaVt) framework which can waive off the requirements of low bandwidth. This is done by transmitting important content only to the end user. The framework is illustrated with the example of video streaming in the context of remote laboratory setup. A framework for distributed processing using mobile agents is discussed with the example of Distributed Video Surveillance (DVS). In this regard, the increased computational costs due to video processing tasks like object segmentation and tracking are shared by the cameras and a local base station called as Processing Proxy Server (PPS).However, in a distributed scenario like traffic surveillance, where moving objects is tracked using multiple cameras, the processing tasks needs to be dynamically distributed. This is done intelligently using mobile agents by migrating from one PPS to another for tracking an individual case object and transmitting required information to the end users. Although the authors propose a specific implementation for CbNaVt and DVS systems, the general ideas in design of such systems exemplify the way information can be intelligently transmitted in any ubiquitous multimedia applications along with the use of mobile agents for real-time processing and retrieval of video signal.

scholarly journals Deep Learning Based Vehicle Detection and Classification Methodology Using Strain Sensors under Bridge Deck

Sensors ◽  
2020 ◽  
Vol 20 (18) ◽  
pp. 5051
Author(s):  
Rujin Ma ◽  
Zhen Zhang ◽  
Yiqing Dong ◽  
Yue Pan
Keyword(s):  
Neural Network ◽  
Bridge Deck ◽  
Vehicle Detection ◽  
Traffic Monitoring ◽  
Transportation Systems ◽  
Strain Sensors ◽  
Long Span ◽  
Long Span Bridge ◽  
Pavement Evaluation ◽  
Close Range

Vehicle detection and classification have become important tasks for traffic monitoring, transportation management and pavement evaluation. Nowadays there are sensors to detect and classify the vehicles on road. However, on one hand, most sensors rely on direct contact measurement to detect the vehicles, which have to interrupt the traffic. On the other hand, complex road scenes produce much noise to consider when to process the signals. In this paper, a data-driven methodology for the detection and classification of vehicles using strain data is proposed. The sensors are well arranged under the bridge deck without traffic interruption. Next, a cascade pre-processing method is applied for vehicle detection to eliminate in-situ noise. Then, a neural network model is trained to identify the close-range following vehicles and separate them by Non-Maximum Suppression. Finally, a deep convolutional neural network is designed and trained to identify the vehicle types based on the axle group. The methodology was applied in a long-span bridge. Three strain sensors were installed beneath the bridge deck for a week. High robustness and accuracy were obtained by these algorithms. The methodology proposed in this paper is an adaptive and promising method for vehicle detection and classification under complex noise. It would serve as a supplement to current transportation systems and provide reliable data for management and decision-making.

scholarly journals The One-Stage Detector Algorithm Based on Background Prediction and Group Normalization for Vehicle Detection

2020 ◽  
Vol 10 (17) ◽  
pp. 5883
Author(s):  
Fei Lu ◽  
Fei Xie ◽  
Shibin Shen ◽  
Jiquan Yang ◽  
Jing Zhao ◽  
...  
Keyword(s):  
Intelligent Transportation Systems ◽  
Network Models ◽  
Vehicle Detection ◽  
Traffic Monitoring ◽  
Transportation Systems ◽  
Neural Network Models ◽  
Detector Performance ◽  
One Stage ◽  
Complex Scenes ◽  
The One

Vehicle detection in intelligent transportation systems (ITS) is a very important and challenging task in traffic monitoring. The difficulty of this task is to accurately locate and classify relatively small vehicles in complex scenes. To solve these problems, this paper proposes a modified one-stage detector based on background prediction and group normalization to realize real-time and accurate detection of traffic vehicles. The one-stage detector firstly adds a module to adjust the width and height of anchors and predict the target background, which avoids the problem of the target vehicle missing detection or wrong detection due to the influence of the complicated environments. Then, group normalization and the loss function based on weight attenuation can improve the one-stage detector performance in the training process. The experimental results on traffic monitoring datasets indicate that the improved one-stage detector is superior to the other neural network models in terms of precision at 95.78%.

Sign in / Sign up

Export Citation Format

Share Document