scholarly journals Frequency Maps as Expert Instructions to Lessen Data Dependency on Real-time Traffic Light Recognition

2020 ◽  
Author(s):  
Hendrik Macedo ◽  
Thiago Almeida ◽  
Leonardo Matos ◽  
Bruno Prado
Keyword(s):  
Real Time ◽  
Autonomous Vehicles ◽  
Machine Learning Algorithms ◽  
Support Vector ◽  
Background Suppression ◽  
Processing Unit ◽  
Traffic Light ◽  
Traffic Lights ◽  
Real Time Traffic ◽  
Computational Resources

Research on Traffic Light Recognition (TLR) has grown in recent years, primarily driven by the growing interest in autonomous vehicles development. Machine Learning algorithms have been widely used to that purpose. Mainstream approaches, however, require large amount of data to properly work, and as a consequence, a lot of computational resources. In this paper we propose the use of Expert Instruction (IE) as a mechanism to reduce the amount of data required to provide accurate ML models for TLR. Given an image of the exterior scene taken from the inside of the vehicle, we stand the hypothesis that the picture of a traffic light is more likely to appear in the central and upper regions of the image. Frequency Maps of traffic light location were thus constructed to confirm this hypothesis. The frequency maps are the result of a manual effort of human experts in annotating each image with the coordinates of the region where the traffic light appears. Results show that EI increased the accuracy obtained by the classification algorithm in two different image datasets by at least 15%. Evaluation rates achieved by the inclusion of EI were also higher in further experiments, including traffic light detection followed by classification by the trained algorithm. The inclusion of EI in the PCANet achieved a precision of 83% and recall of 73% against 75.3% and 51.1%, respectively, of its counterpart. We finally presents a prototype of a TLR Device with that expert model embedded to assist drivers. The TLR uses a smartphone as a camera and processing unit. To show the feasibility of the apparatus, a dataset was obtained in real time usage and tested in an Adaptive Background Suppression Filter (AdaBSF) and Support Vector Machines (SVMs) algorithm to detect and recognize traffic lights. Results show precision of 100% and recall of 65%.

scholarly journals Real Time Traffic Light Detection by Autonomous Vehicles using Artificial Neural Network Techniques

2019 ◽  
Vol 8 (10) ◽  
pp. 2129-2133
Keyword(s):  
Neural Network ◽  
Real Time ◽  
Autonomous Vehicles ◽  
Traffic Light ◽  
Artificial Intelligence Technique ◽  
Light Detection ◽  
Traffic Lights ◽  
Real Time Traffic ◽  
The Right ◽  
Sufficient Precision

Autonomous vehicles are the reality of the future, they will open up the way for future advanced systems where computers are expected to take over the decision making of driving. These automobiles are capable of sensing their environment and moving with little or no human input. The main goal of this research is to detect traffic light in real-time for autonomous vehicles. Apart from taking decisions to navigate in the right manner the autonomous vehicles important task is to detect traffic lights, so that it can obey the traffic rules with sufficient precision. The work carried out in this research makes use of two Artificial Intelligence technique, these techniques are compared in accomplishing the task of traffic light detection in real time. The two models that are designed and implemented are Convolution neural network (CNN) and Deep Convolution Inverse Graphics Network (DCIGN). The results clearly show that DCIGN out performance CNN by 8%.

scholarly journals Prototyping a Traffic Light Recognition Device with Expert Knowledge

Information ◽  
2018 ◽  
Vol 9 (11) ◽  
pp. 278 ◽  
Author(s):  
Thiago Almeida ◽  
Hendrik Macedo ◽  
Leonardo Matos ◽  
Nathanael Vasconcelos
Keyword(s):  
Expert Knowledge ◽  
Support Vector ◽  
Background Suppression ◽  
Processing Unit ◽  
Driver Assistance ◽  
Traffic Light ◽  
Accuracy Rate ◽  
Light Detection ◽  
Vector Machines ◽  
Detection And Recognition

Traffic light detection and recognition (TLR) research has grown every year. In addition, Machine Learning (ML) has been largely used not only in traffic light research but in every field where it is useful and possible to generalize data and automatize human behavior. ML algorithms require a large amount of data to work properly and, thus, a lot of computational power is required to analyze the data. We argue that expert knowledge should be used to decrease the burden of collecting a huge amount of data for ML tasks. In this paper, we show how such kind of knowledge was used to reduce the amount of data and improve the accuracy rate for traffic light detection and recognition. Results show an improvement in the accuracy rate around 15%. The paper also proposes a TLR device prototype using both camera and processing unit of a smartphone which can be used as a driver assistance. To validate such layout prototype, a dataset was built and used to test an ML model based on adaptive background suppression filter (AdaBSF) and Support Vector Machines (SVMs). Results show 100% precision rate and recall of 65%.

scholarly journals A Proposal of Implementation of Sitting Posture Monitoring System for Wheelchair Utilizing Machine Learning Methods

Sensors ◽  
2021 ◽  
Vol 21 (19) ◽  
pp. 6349
Author(s):  
Jawad Ahmad ◽  
Johan Sidén ◽  
Henrik Andersson
Keyword(s):  
Machine Learning ◽  
Pressure Distribution ◽  
Real Time ◽  
Monitoring System ◽  
Pressure Ulcers ◽  
Machine Learning Algorithms ◽  
Raspberry Pi ◽  
Support Vector ◽  
Processing Unit ◽  
Posture Recognition

This paper presents a posture recognition system aimed at detecting sitting postures of a wheelchair user. The main goals of the proposed system are to identify and inform irregular and improper posture to prevent sitting-related health issues such as pressure ulcers, with the potential that it could also be used for individuals without mobility issues. In the proposed monitoring system, an array of 16 screen printed pressure sensor units was employed to obtain pressure data, which are sampled and processed in real-time using read-out electronics. The posture recognition was performed for four sitting positions: right-, left-, forward- and backward leaning based on k-nearest neighbors (k-NN), support vector machines (SVM), random forest (RF), decision tree (DT) and LightGBM machine learning algorithms. As a result, a posture classification accuracy of up to 99.03 percent can be achieved. Experimental studies illustrate that the system can provide real-time pressure distribution value in the form of a pressure map on a standard PC and also on a raspberry pi system equipped with a touchscreen monitor. The stored pressure distribution data can later be shared with healthcare professionals so that abnormalities in sitting patterns can be identified by employing a post-processing unit. The proposed system could be used for risk assessments related to pressure ulcers. It may be served as a benchmark by recording and identifying individuals’ sitting patterns and the possibility of being realized as a lightweight portable health monitoring device.

A REAL TIME TRAFFIC LIGHT RECOGNITION SYSTEM

2008 ◽  
Vol 05 (02) ◽  
pp. 149-161 ◽  
Author(s):  
XIAOLI YANG ◽  
ZHENPENG ZHAO ◽  
YOUN K. KIM
Keyword(s):  
Real Time ◽  
Shape Analysis ◽  
Feature Matching ◽  
Median Filter ◽  
Recognition System ◽  
Transport Systems ◽  
Traffic Light ◽  
Traffic Lights ◽  
Real Time Traffic ◽  
Color Detection

Driving is a challenging task, especially for those with color vision deficiencies. Intelligent Transport Systems (ITS) can provide useful information and make driving safe and convenient. A real time traffic light recognition system is presented in this paper for improving public safety and facilitating color deficient drivers. It may also be included as a part of ITS. The presented system consists of a digital video camera to record traffic lights and a portable PC to process images in real time. It uses various techniques of color detection, feature matching with normalized cross-correlation (NCC), and mathematical shape analysis for the traffic light recognition. For color detection, we obtained an initial solution by using RGB component adjustment, thresholding algorithm, and median filter. In dealing with illumination changes with weather and time, a simple adaptation method was developed. Feature matching with NCC was used after color detection to further detect and recognize the traffic lights. To improve the system's tolerance and robustness, a mathematical shape analysis was undertaken to obtain the final results. Numerous experiments were conducted to demonstrate the effectiveness and practicability of the system with images under different weather conditions. The average recognition ratio is higher than 95% from the testing results. The average processing time is 30 ms per frame, making the system suitable in real time conditions. Audio alert is added to the current system as an integral part of a portable system to be developed.

scholarly journals Real-Time Traffic Light Detection and Interpretation using Circle Centroid

2019 ◽  
Vol 8 (12S2) ◽  
pp. 465-469
Keyword(s):  
Real Time ◽  
Hough Transform ◽  
Weather Condition ◽  
Traffic Light ◽  
Image Processing Algorithm ◽  
Traffic Lights ◽  
Real Time Traffic ◽  
Yellow Color ◽  
Red Color ◽  
Different Color

The paper proposes a new method to recognize the sequence of a traffic light using image processing algorithm. Invariant in factor lightning and weather condition that lead to misinterpret the color of traffic light is one of the factors of accident at traffic light conjunction besides the behavior of the driver itself. The process to identify the color and shape of traffic light are Image Acquisition, Pre-Processing, Detection, Feature Extraction and Interpretation. RGB normalization is performed and simple thresholding method that acts as color segmentation provides a better division of the traffic light colors. Circle Hough Transform and HSV color features based on the traffic light aspect are used to decide whether the spots on the frames are likely to be traffic lights’ color and shape. The detection of traffic light will be obtained after identifying the feature such as the centroid of the Circle Hough Transform that need to be extracted at the end of the result. The research has been improved by focusing on detection and interpretation of traffic light based on real time video rather than image sample as the input. The proposed algorithm can detect the green color accurately with maximum accuracy of 83.8%, yellow color of 75.6% and red color with minimum accuracy of 70.19%. This indicates that there is a possibility to use the proposed algorithm to detect the three different color of green, yellow and red color of traffic lights’ colour.

scholarly journals Mobile Road Traffic Management System Using Weighted Sensor

2017 ◽  
Vol 11 (5) ◽  
pp. 147 ◽  
Author(s):  
Solomon Adegbenro Akinboro ◽  
Johnson A Adeyiga ◽  
Adebayo Omotosho ◽  
Akinwale O Akinwumi
Keyword(s):  
Real Time ◽  
Traffic Control ◽  
Urban Areas ◽  
Traffic Management ◽  
Management System ◽  
Road Traffic ◽  
Traffic Information ◽  
Traffic Light ◽  
Real Time Traffic ◽  
Traffic Management System

<p><strong>Vehicular traffic is continuously increasing around the world, especially in urban areas, and the resulting congestion ha</strong><strong>s</strong><strong> be</strong><strong>come</strong><strong> a major concern to automobile users. The popular static electric traffic light controlling system can no longer sufficiently manage the traffic volume in large cities where real time traffic control is paramount to deciding best route. The proposed mobile traffic management system provides users with traffic information on congested roads using weighted sensors. A prototype of the system was implemented using Java SE Development Kit 8 and Google map. The model </strong><strong>was</strong><strong> simulated and the performance was </strong><strong>assessed</strong><strong> using response time, delay and throughput. Results showed that</strong><strong>,</strong><strong> mobile devices are capable of assisting road users’ in faster decision making by providing real-time traffic information and recommending alternative routes.</strong></p>

scholarly journals A Novel Fast Training Method for SVM and Its Application in Fault Diagnosis of Service Robot

2015 ◽  
Vol 11 (6) ◽  
pp. 4 ◽  
Author(s):  
Xianfeng Yuan ◽  
Mumin Song ◽  
Fengyu Zhou ◽  
Yugang Wang ◽  
Zhumin Chen
Keyword(s):  
Fault Diagnosis ◽  
Classification Accuracy ◽  
Machine Learning Algorithms ◽  
Sensor Data ◽  
Service Robot ◽  
Support Vector ◽  
Svm Classifier ◽  
Processing Unit ◽  
Training Method ◽  
Fast Training

Support Vector Machines (SVM) is a set of popular machine learning algorithms which have been successfully applied in diverse aspects, but for large training data sets the processing time and computational costs are prohibitive. This paper presents a novel fast training method for SVM, which is applied in the fault diagnosis of service robot. Firstly, sensor data are sampled under different running conditions of the robot and those samples are divided as training sets and testing sets. Secondly, the sampled data are preprocessed and the principal component analysis (PCA) model is established for fault feature extraction. Thirdly, the feature vectors are used to train the SVM classifier, which achieves the fault diagnosis of the robot. To speed up the training process of SVM, on the one hand, sample reduction is done using the proposed support vectors selection (SVS) algorithm, which can ensure good classification accuracy and generalization capability. On the other hand, we take advantage of the excellent parallel computing abilities of Graphics Processing Unit (GPU) to pre-calculate the kernel matrix, which avoids the recalculation during the cross validation process. Experimental results illustrate that the proposed method can significantly reduce the training time without decreasing the classification accuracy.

Sign in / Sign up

Export Citation Format

Share Document