Search

Variable-message signs, which are an important type of traffic sign, have been widely used in traffic control. This study examines the influence of text luminance, text colour and background luminance of variable-message signs on legibility in urban areas under night-time vision. First, the background luminance of variable-message signs in typical urban areas is investigated, tested and classified as high, medium or low depending on the obtained values (3.0, 1.0 or 0.1 cd/m2, respectively). Second, a calculation model is established for the experimental visual recognition distance under different visual acuity conditions, the experimental luminance interval is determined based on psychophysical magnitude perceived, and a legibility experiment is performed using Landolt-C rings. Finally, the reasonable luminance interval (minimum [replacement] luminance), optimal luminance and changing trends of luminance for red, green and yellow colours under different background luminances are obtained such that the legibility requirement is satisfied. The results indicated that different colours have different luminance interval and optimal luminance under different background luminances. This study is of great importance in determining the reasonable luminance of variable-message sign texts and improving the safety and efficiency of road traffic. In the future, these research results are expected to serve as guidelines for variable-message sign setup and management.

The state-of-the-art traffic sign recognition (TSR) algorithms are designed to recognize the textual information of a traffic sign at over 95% accuracy. Even though, they are still not ready for complex roadworks near ramps. In real-world applications, when the vehicles are running on the freeway, they may misdetect the traffic signs for the ramp, which will become inaccurate feedback to the autonomous driving applications and result in unexpected speed reduction. The misdetection problems have drawn minimal attention in recent TSR studies. In this paper, it is proposed that the existing TSR studies should transform from the point-based sign recognition to path-based sign learning. In the proposed pipeline, the confidence of the TSR observations from normal vehicles can be increased by clustering and location adjustment. A supervised learning model is employed to classify the clustered learned signs and complement their path information. Test drives are conducted in 12 European countries to calibrate the models and validate the path information of the learned sign. After model implementation, the path accuracy over 1,000 learned signs can be increased from 75.04% to 89.80%. This study proves the necessity of the path-based TSR studies near freeway ramps and the proposed pipeline demonstrates a good utility and broad applicability for sensor-based autonomous vehicle applications.

Accurate understanding of traffic signs is integral to driving safety. Questionnaires and behavior tests are usually used to measure sign comprehension. However, biometric indicators have also been developed to measure sign comprehension. This study investigated neural indicators underlying traffic sign understanding by measuring event-related potentials (ERPs). A stimulus 1–stimulus 2 paradigm was adopted. A sign identification experiment was conducted independently in a laboratory setting with 60 participants divided into two groups. ERPs and brain topographies were analyzed to evaluate traffic sign comprehension. Misunderstanding was associated with greater amplitudes of the N300 and N400. The study demonstrates that the N300 and N400 are effective neural indicators for evaluating sign comprehension and can be used for predicting drivers’ sign comprehension and behavior.

Work zones affect traffic safety and efficiency by changing the road condition and drivers’ maneuver. Therefore, it is important to fully understand drivers’ merging behavior in work zone related areas. In this study, a model framework composed of decision-distance analysis and merging-distance analysis was proposed, which could describe both decision-making and lane-changing process of merging behavior. A road environment with work zone was developed based on a driving simulator, and six scenarios composed of two levels of traffic volume and three different lane-end sign's locations were designed. Forty-two licensed participants, including 21 females (10 professional drivers vs. 11 normal drivers) and 21 males (15 professional drivers vs. 6 normal drivers) finally completed the experiment. Based on the experimental data, parametric survival models were established to analyze the effects of traffic sign location, traffic situation and driver characteristics on drivers’ decision distance and merging distance. The results showed that: (i) the lane-end sign's location affected the decision point of lane changing and further affected the merging distance. However, the effect was weakened when the sign was placed far away from the work zone; (ii) merging distance in high traffic volume condition was shorter than that in low traffic volume condition; (iii) work zone posed greater challenges for female drivers as they merged later than males, and females were reluctant to adjust their merging distance according to different gap conditions. The findings shed some light on the future improvement of traffic design and management in work zones.

Highway navigation is often affected by complex topography, and the flat curve plays an important role in the horizontal alignment design of a highway. Many curves are formed, where visibility could be decreased. Thus, the indicative function of a traffic sign plays a crucial role in ensuring driving safety at the curve. Due to the blocked visibility, the probability of the traffic sign occlusion at the curve of operating highways is quite high. It is urgent to consider the clearing obstructions around traffic signs at curves during highway construction. In this study, the potential of visual occlusion for traffic signs on curved highways was investigated. Firstly, the driver’s visibility window that contains traffic signs was defined and criteria of visual occlusion were proposed. Secondly, a geometric occlusion design formula was established to mimic the visual recognition process of traffic signs on a curved highway, yielding the formula to calculate the visibility window. Finally, the occlusion design formula was applied into a case study of the Beijing-Hong Kong-Macau Expressway (Hunan section), in which visibility windows were calculated and analyzed. The obtained results verified the correctness and effectiveness of the occlusion design formula developed in this study.

The road environment recognition using embedded technologies has been researched intensively recently. Several papers deal with the detection of the urban road environment-type (RET), such as downtown, residential area, and business/industrial area. These RETs can characterize the road environment around an ego-car. A RET detection approach taking into account relevant traffic signs (TSs) that are visible from the ego-car – along its route – was also proposed. It was assumed that the TS data, namely the type and the location of each detected TS along the route, was made available for the purpose by an on-board TS recognition system (TSR). The TS data is constantly updated, aggregated and evaluated in a multi-scale manner by a RET detection system, so one can produce a probability series of occurrence of each TS type with respect to each of the considered urban RETs. In the present paper, we develop a heuristic for dynamic RET detection using fuzzy similarities on a special graph. We also propose the generic process of this approach, which will be the subject of further development and testing.

Traffic sign detection is one of the critical technologies in the field of intelligent transportation systems (ITS). The difficulty of traffic sign detection mainly lies in detecting small objects in a wide and complex traffic scene quickly and accurately. In this paper, we regard traffic sign detection as a region classification problem and propose a two-stage CNN-based approach to solve it. At the first stage, we design an efficient network which is built with improved fire-modules to generate object proposals quickly. The network up-samples and merges the feature maps of different scales to attain a high-resolution fused feature map which contains semantically strong features of multi-scale objects. Specially, the prediction is made on the fuse feature map and based on the novel center-point estimation. With the overall designs, our region proposal network can achieve high recall value while using low-resolution images. At the second stage, a separate classification network is proposed. The bottleneck of the classification performance is generally caused by the greatly similar appearances between traffic signs. Therefore, we further explore local regions with critical differences between traffic signs to obtain fine-grained local features which help to improve classification. Finally, we evaluate our method on a challenge benchmark Tsinghua-Tencent 100K which provides many large images with small traffic sign instances. The experiment result shows that our method has better performance and faster detection speed than many state-of-the-art traffic sign detection methods.

Nowadays, the number of road accident in Malaysia is increasing expeditiously. One of the ways to reduce the number of road accident is through the development of the advanced driving assistance system (ADAS) by professional engineers. Several ADAS system has been proposed by taking into consideration the delay tolerance and the accuracy of the system itself. In this work, a traffic sign recognition system has been developed to increase the safety of the road users by installing the system inside the car for driver’s awareness. TensorFlow algorithm has been considered in this work for object recognition through machine learning due to its high accuracy. The algorithm is embedded in the Raspberry Pi 3 for processing and analysis to detect the traffic sign from the real-time video recording from Raspberry Pi camera NoIR. This work aims to study the accuracy, delay and reliability of the developed system using a Raspberry Pi 3 processor considering several scenarios related to the state of the environment and the condition of the traffic signs. A real-time testbed implementation has been conducted considering twenty different traffic signs and the results show that the system has more than 90% accuracy and is reliable with an acceptable delay.

I detta arbete utvecklas ett nytt stöldförebyggande skyltsystem för tillfälliga trafikskyltar. Fästet hos det utvecklade skyltsystemet bidrar till att stöld och vandalisering av skyltarna försvåras genom att vara låsande. Viktig trafikinformation kan därför bevaras i trafiken. Det nya skyltsystemet möjliggör montering av flera skyltar på båda sidor av stolpen med steglös höjdjustering. Skyltsystemet som utvecklats har många av det befintliga systemets fördelar. Det har även fördelar som det befintliga systemet saknar och stor utvecklingspotential. I arbetet utförs även simulationer, beräkningar, hållfasthetstester och användartester för att säkerställa att systemet är säkert att använda i trafiken. In this project, a new anti-theft sign system is being developed for temporary traffic signs. The connection between the sign and pole of the developed sign system contributes to the theft and vandalism of the signs being made more difficult by being self-locking. Important traffic information can therefore remain in traffic. The new sign system enables mounting of several signs on both sides of the post with stepless height adjustment. The sign system that has been developed has many of the existing systems advantages. It also has advantages that the existing system lacks and great development potential. The project also contains simulations, calculations, strength tests and user tests to ensure that the system is safe to use in traffic.

The legibility of traffic signs has been considered from the beginning of design, and traffic signs are easy to identify for humans. For computer systems, however, identifying traffic signs still poses a challenging problem. Both image-processing and machine-learning algorithms are constantly improving, aimed at better solving this problem. However, with a dramatic increase in the number of traffic signs, labelling a large amount of training data means high cost. Therefore, how to use a small number of labelled traffic sign data reasonably to build an efficient and high-quality traffic sign recognition (TSR) model in the Internet-of-things–based (IOT-based) transport system has been an urgent research goal. Here, the authors propose a novel semi-supervised learning approach combining global and local features for TSR in an IOT-based transport system. In their approach, histograms of oriented gradient, colour histograms (CH), and edge features (EF) are used to build different feature spaces. Meanwhile, on the unlabelled samples, a fusion feature space is found to alleviate the differences between different feature spaces. Extensive evaluations on a collection of signs from the German Traffic Sign Recognition Benchmark (GTSRB) dataset shows that the proposed approach outperforms the others and provides a potential solution for practical applications.

Aiming at recognizing small-scale and complex traffic signs in the driving environment, a traffic sign detection algorithm YOLO-FAM based on YOLOv5 is proposed. Firstly, a new backbone network, ShuffleNet-v2, is used to reduce the algorithm’s parameters, realize lightweight detection, and improve detection speed. Secondly, the Bidirectional Feature Pyramid Network (BiFPN) structure is introduced to capture multi-scale context information, so as to obtain more feature information and improve detection accuracy. Finally, location information is added to the channel attention using the Coordinated Attention (CA) mechanism, thus enhancing the feature expression. The experimental results show that compared with YOLOv5, the mAP value of this method increased by 2.27%. Our approach can be effectively applied to recognizing traffic signs in complex scenes. At road intersections, traffic planners can better plan traffic and avoid traffic jams.

Traffic sign recognition (TSR) is a key aspect involved in the development of robust automated transportation systems. It inherently involves the task of traffic sign detection (TSD), which can be challenging due to traffic signs often being subject to deterioration or occlusion, caused by various environmental factors, or through actions of vandalism. Even though, notable advancements have been achieved in the areas of TSR and TSD, few studies have provided robust algorithms, able to be generalized in real-world applications. This mostly stems from the lack of an extensive traffic sign dataset, standardized for benchmarking purposes. In light of the aforementioned, this paper presents a novel traffic sign dataset, which consists of the Carla Traffic Sign Detection (CTSD), and the Carla Traffic Sign Recognition Dataset (CATERED), targeting the detection and recognition processes respectively. Using the proposed dataset for training and evaluation, a deep Auto-Encoder algorithm is presented, demonstrating high accuracy in detecting and recognizing the distorted traffic signs. Finally, the system is further extended to a federated learning environment, exemplifying its applicability in modern decentralized and interconnected architectures.

One of the intelligent transportation system's critical tasks is to understand traffic signs and convey traffic information to humans. However, most related works are focused on the detection and recognition of traffic sign texts or symbols, which is not sufficient for understanding. Besides, there has been no public dataset for traffic sign understanding research. Our work takes the first step towards addressing this problem. First, we propose a "CASIA-Tencent Chinese Traffic Sign Understanding Dataset" (CTSU Dataset), which contains 5000 images of traffic signs with rich semantic descriptions. Second, we introduce a novel multi-task learning architecture that extracts text and symbol information from traffic signs, reasons the relationship between texts and symbols, classifies signs into different categories, and finally, composes the descriptions of the signs. Experiments show that the task of traffic sign understanding is achievable, and our architecture demonstrates state-of-the-art and superior performance. The CTSU Dataset is available at http://www.nlpr.ia.ac.cn/databases/CASIA-Tencent%20CTSU/index.html.

Traffic sign board recognition is a very significant work for the upcoming driver assistance intelligent vehicle systems. The ability to detect such traffic signs from the real road scenes intensifies the safety of the intelligent vehicle systems. However, automatic detection and classification of traffic signs by such intelligent vehicle systems is a challenging task due to various factors such as variation in light illumination, different viewpoints, colour faded traffic sign, motion blurring, etc. The deep learning models have proved to provide solutions to overcome these factors. This paper proposed deep learning-based residual network for traffic sign detection and classification (DLRN-TSDC) model for effective Indian Traffic Sign Board Recognition. The DLRN-TSDC model makes use of Colour space threshold segmentation technique for the effective identification of sign boards. Simultaneously, pre-processing of the detected traffic sign takes place in three distinct ways such as clipping of edges, image enhancement and size normalization. In addition, the ResNet-50 model is used as a feature extractor and a classifier to determine the final class label of the traffic sign board. Extensive experimental analysis was carried out to validate the effective performance of the DLRN-TSDC model and for the precision, recall, Intersection over Union (IoU) and accuracy scores are 98.76%, 98.92%, 89.56% and 98.84%, respectively.

The developments of computer aided driving systems detection and classification of traffic signs getting moreimportant everyday. In this subject there has been many different methods proposed and used. The morphological features of trafficsigns carries a lot of significance in terms of classification. The main idea of this study is using morphological features in deep learning network to enhance classification success rate. In this study the use of deep morphological networks is proposed for the sake of better utilizing morphological features in traffic sign efficiently.The results of deep morphological networks and convolutional neural networks are compared and deep morphological networks shows higher classification success rate.

Traffic sign recognition is a well-researched problem in computer vision. However, the state of the art methods works only for frequent sign classes, which are well represented in training datasets. We consider the task of rare traffic sign detection and classification. We aim to solve that problem by using synthetic training data. Such training data is obtained by embedding synthetic images of signs in the real photos. We propose three methods for making synthetic signs consistent with a scene in appearance. These methods are based on modern generative adversarial network (GAN) architectures. Our proposed methods allow realistic embedding of rare traffic sign classes that are absent in the training set. We adapt a variational autoencoder for sampling plausible locations of new traffic signs in images. We demonstrate that using a mixture of our synthetic data with real data improves the accuracy of both classifier and detector., Comment: The paper was submitted to the journal "Computer Optics" and is currently under review

Traffic sign and light detection are core components of Advanced Driver Assistance Systems (ADAS) and self-driving vehicles. The automotive industry is widely employing numerous approaches for automation through computer vision techniques. Object detection algorithms based on deep learning can be divided into two main categories, two stage and single stage detection algorithms. Two stage algorithms are designed to improve detection accuracy. While single stage algorithms are constructed to be faster, this increases their suitability for real time applications. This thesis presents a lightweight traffic sign and light detector by adapting a single stage, Single Shot Multibox Detection (SSD) algorithm by providing both high accuracy and real time detection capability. Therefore, the Visual geometry group (VGG16) base network in original SSD is replaced by MobileNet, that expertly manages detection speed and network size because of its lighter architecture. It is essential for the application domain to be able to detect small objects which is what the original SSD struggles with. For autonomous driving the detection results with respect to the distance of an object is of particular interest. A comfortable braking distance is needed in case of traffic signs and lights. That requires object detection from a farther distance, but farther distance makes the object to be detected appear smaller. Thus, this work further optimizes the number of feature map layers of the algorithm for the detection of small objects along with a better trade off between accuracy and detection time. Experimental results confirm the effectiveness of the proposed model as compared to the standard SSD with VGG16 and SSD with MobileNet V2.

Due to the expeditious increase in the number of vehicles, there is an increase in the number of road casualties even in a highly sophisticated roadway. This depicts the natural limitation of a human in maintaining Traffic rules. To avoid any lethal circumstance assistive driving vehicles are introduced which consists of systems that guide drivers in different Traffic situations. Traffic sign recognition systems play a crucial role in assistive driving vehicles these systems have been based on characteristics of the sign and two-state detectors due to accuracy and real-time factors systems on these bases are not used for real-time application. In this paper, we present a system that can recognize Traffic signs and their distance from the vehicle in non-ideal lighting as well as in varying climatic conditions. Our work proceeds with the implementation of YOLOv3(deep convolutional network based on end-to-end detection algorithm) used for Traffic sign recognition and segmentation. Training of the model is done with GTSRB dataset and achieves an accuracy of about 98.5% for the recognition task in different real-time scenarios. Furthermore, an efficient Heuristic-based approach has been deployed for estimating the distance between the Traffic sign and the monocular camera(placed in the vehicle) at every instance.

This paper presents an approach to detect traffic signs using You Only Look Once version 4 (YOLOv4) model. The traffic sign detection and recognition system (TSDR) play an essential role in the intelligent transportation system (ITS). TSDR can be utilized for driver assistance and, eventually, driverless cars to reduce accidents. When driving an automobile, the driver's attention is usually drawn to the road. On the other hand, most traffic signs are situated on the side of the road, which may have contributed to the collision. TSDR allows drivers to view traffic sign information without having to divert their attention. Due to the existence of a large background, clutter, fluctuating degrees of illumination, varying sizes of traffic signs, and changing weather conditions, TSDR is an important but difficult process in intelligent transport systems. Many efforts have been made to find answers to the major issues that they face. The objective of this study addresses road traffic sign detection and recognition using a technique that initially detects the bounding box of a traffic sign. Then the detected traffic sign will be recognized for usage in a speeded-up process. Since safe driving necessitates real-time traffic sign detection, the YOLOv4 network was employed in this research. YOLOv4 was evaluated on our dataset, which consisted of manual annotations to identify 43 distinctive traffic signs classes. It was able to achieve an average recognition accuracy of 84.7%. Overall, the work adds by presenting a basic yet effective model for real-time detection and recognition of traffic signs.

Traffic sign detection and recognition (TSDR) is one of the main area of research in autonomous vehicles and Advanced Driving Assistance System (ADAS). In this paper a method is proposed to detect and classify the prohibitory subset of German Traffic Sign data set. The traffic sign detection module utilizes adaptive color segmentation based on mean saturation value of local neighborhood, and Circular Hough Transform (CHT) to locate the traffic signs in the input images. Adaptive color thresholding shows improvement in segmenting traffic signs where images have uneven lighting or very high or low contrast levels, compared to global thresholding. Furthermore, number of false alarms are minimized by utilizing an additional validation stage. For the recognition phase of the algorithm, multiple deep Convolutional Neural Networks (CNN) with different structures are developed from scratch to compare their performance and identify the network with highest accuracy.

Self driving cars are picking up pace and thus serves as a challenging research domain. One of the most crucial functions in an autonomous vehicle is accurately detecting and recognising the traffic signs. Traffic sign detection is a crucial task in traffic sign recognition systems. Deep neural networks are proven powerful in traffic sign classification. As traffic sign violation leads to law and order disruption, posing a threat to human life, there is a need for robust algorithms with efficient actuation to perform the delegated task. Thus, this paper proposes a robust algorithm which addresses challenges like haze, fog, unclear images due to improper condition of roads to efficiently detect and recognize traffic signs using Image manipulation, Optical Character Recognition (OCR) algorithm and You Only Look Once (YOLOv3) detection algorithm. The proposed algorithm in this paper for hazy images that aids in ADAS applications has an accuracy of 87.29%. This contributes to the emerging research domains of autonomous vehicles and self-driving cars.

This paper proposes a framework that will detect different types of traffic signs from images for Traffic sign recognition and detection in advanced driver support system. In this paper classical LeNet5 network model is used with GTSRB dataset for evaluation purpose. Shape and color are the two main factors that distinguish road signs from other objects. Firstly, preprocessing of an image to highlight important information. Secondly, the training data are scaled, normalized, expanded, and augmented. Finally, the validation and testing phase in which traffic signs are classified and accuracy is calculated. It is concluded from the results that our system has achieved greater accuracy due to the architectural optimization of the neural network for detecting traffic signs. Significantly better performance was achieved with partially blurred images. Traffic sign are classified in 43 classes. For training 22271 images are used in which 6960 images used for testing and 5568 images are used in the evaluation purpose. With new design model 98.75% accuracy is obtained.

In order to observe and recognize road signs, it is necessary to prepare a system that is mainly based on visibility. The traffic sign recognition system collects traffic sign data and allows the driver. Traffic signs from a series of images with inaccurate color data. Color-based unit area segmentation techniques are used to detect traffic signs. To improve the targeting performance, it is highly required to use products with the highest hue and saturation coefficients. To achieve higher shape classification performance, a linear support vector machine with raster space options are used for fractional blobs. Area unit recognition of traffic signs is applied to non-linear support vector multiple classifiers with the desired pixels that are connected to edges.

Traffic signs play a critical role in the safety and efficiency of any roadway; however, limited information exists on current traffic sign inventories (TSIs). The size of current traffic s...

The Advanced Driver Assistance System includes traffic sign identification and recognition. Traffic signs warn drivers of traffic laws, road conditions, and route directions, assisting them in driving more efficiently and safely. Traffic Sign Recognition is a technique for regulating traffic signals, warning drivers, and commanding or prohibiting specific acts. A quick real-time and reliable automated traffic sign detection and recognition system can assist and relieve the driver, improving driving safety and comfort significantly. For autonomous intelligent driving vehicles or driver assistance systems, automatic identification of traffic signals is also important. This paper aims to use Neural Networks to identify traffic sign patterns. Several image processing methods are used to pre-process the images. Then, to understand traffic sign patterns, Neural Networks stages are performed. To find the best network architecture, the system is trained and validated. The results of the experiments show that traffic sign patterns with complex backgrounds can be classified very accurately.

In today’s technology-driven world, there are a lot of devices and products have been invented including the Autonomous Vehicle (AV). Autonomous vehicle is a driverless vehicle which able to drive on its own. It works with the aid of various embedded systems and sensors that help the AV to function greatly. Speed limit signs recognition is one of the TSR essential features for AV where it helps to automatically detect and recognize speed limit signs on the road. There are many speed limit signs available on the road including 30km/h, 60km/h, 90km/h signs and to name a few. However, the interclass similarity among these speed limit signs has created a challenge for the TSR system in detection and recognition process. Therefore, this study proposes image processing technique to develop the speed limit sign recognition for TSR system utilizing a single layer network called Single Shot Multibox Detector (SSD) algorithm. The German Traffic Sign Dataset (GTSD) is used for the purpose of training the model and the model is then tested using the real-time images of standard Malaysian speed limit signs. An accuracy testing using confusion matrix is conducted to find the overall accuracy of the system. A total 100 images are used during testing and the system achieved over 92.4% of the average accuracy for detection and recognition of the speed limit signs.

Maintaining a database with the type, location, and direction of traffic signs is a labor-intensive part of asset management for many road authorities. Today there are high-quality cameras in cell-phones that can add location (EXIF) metadata to the images. This makes it efficient and cheap to collect large geo-located imagery datasets. Detecting traffic signs from imagery is also much simpler today due to the availability of several high-quality open-source object-detection solutions. In this paper, we use the detection of traffic signs to find both the location and the direction of physical traffic signs. Five approaches to cluster the detections are presented. An extensive experimental evaluation shows that it is important to consider both the location and the direction. The evaluation is done on a novel dataset with 21,565 images that is available free for download. This includes the ground-Truth location of 277 traffic signs and all source code. The conclusion is that traffic signs are detected with an F1 score of 0.8889, a location accuracy of 5.097-meter (MAE), and a direction accuracy of ± 11.375°(MAE). Only data from two trips are needed to get these results.

The problem related to traffic is a complex one requiring proper design and planning for developing an efficient solution. Drivers are expected to pay attention to identify traffic signs, interpret and follow them while driving. Many of the accidents occur because drivers lose their attention and ignore the street signs. An automatic system in a car which detects, recognizes, interprets and gives warning to the driver would be a great help in reducing the number of road accidents. This paper proposes an embedded system based on Raspberry Pi 3 and uses CNN (Convolution Neural Network) algorithm to detect and identify the captured traffic signs and activate the motor accordingly. Among the different activation functions ReLu is better but its derivative is zero fo r negative values. Hence in this paper a new activation function is proposed which gives a better accuracy, less loss and execution time when compared with ReLu. The detection of traffic sign boards will send the output signal to PIC (Peripheral Interface Controllers) microcontroller which will compare the input value from Raspberry Pi with the rotating speed of DC motor. According to the control signal from PIC microcontroller, servomotor turns 45 deg for controlling ABS system and the speed of the vehicle is thus reduced when a sign board is detected. LCD displays the speed in which the vehicle is running. The CNN model with proposed activation function gives an accuracy of 93% with sensitivity and specificity above 70% and 55% respectively.

Whenever it comes to road safety, traffic signs plays a vital role. Understanding and following different categories of traffic sign is also important. Several methods are available to tackle this problem of traffic sign detection and one of the method is the object detection algorithm. Generic object detectors such as R-CNN, Fast RCNN, Faster RCNN, YOLO, etc. can perform multiple detection in a frame. Generic object detectors fails in identifying small objects with in a frame due to the usage of pooling with the convolutional layer. Pooling layer, filters out the low level features, but this cannot happen when we deal with real time cases. Because in real scenarios the traffic sign contributes only a small portion of the image. TT100K dataset is the traffic sign dataset that can be used to train proposed model as it meets the needs of our problem. Here in this paper we try to overcome the issue of low level features by employing ridgelet filters to get the enhanced features.

Traffic sign recognition is a computer vision technique to recognize the traffic signs put on the road. In this paper, a traffic sign dataset with approximately 5000 images is collected. This paper presents an ablation analysis of Multilayer Perceptron and Convolutional Neural Networks in traffic sign recognition. The ablation analysis studies the effects of different architectures of Multilayer Perceptron and Convolutional Neural Networks, batch normalization, and dropout. A total of 8 different models are reviewed and their performance is studied. The experimental results demonstrate that Convolutional Neural Networks outperform Multilayer Perceptron in general. Leveraging dropout layer and batch normalization is effective in improving the stability of the model and achieved 98.62% accuracy in traffic sign recognition.

Road traffic signs are part of the road infrastructure components meant to ensure the safe use of roads by motorists and pedestrians alike. A good knowledge and comprehension of road signs helps ensure smooth flow of traffic, improves safety for other motorists and provides drivers with good reaction time to comply with the message dictated by the signs. Abu Dhabi, being the capital of the UAE, brings together people from all over the world for work, business or tourism. Abu Dhabi has quality roads and traffic signage together with an efficient police force that enforces adherence to traffic rules. Despite all these and the reduced fatalities in absolute terms, traffic violations have been increasing exponentially, resulting in five million traffic fines in 2019 for a population of about 2.9 million inhabitants. This study sought to assess motorists’ comprehension of the various traffic signs used across the UAE. The results of the survey (N = 200) revealed that the drivers were able to correctly identify 77% of the road signs in the survey. It was also found that the respondents failed to correctly identify traffic signs within the specific category referred to as “advance warning signs”. Respondents’ education, nationality, gender, and marital status did not have any statistically significant effects on the results compared to earlier studies. It is recommended that more attention be focused on driver education and training.

Traffic sign is important as they provide visual information to drivers. Traffic signal detection is a technique used detects traffic signs put on a road and inform it to a driver and it is an integral part of intelligent driver assistance. This paper proposed as hardware accelerator using cascade classifiers based on FPGA to detect the traffic signs. Cascade classifier consists of number of stages and adaptive boosting technique is trainer for the cascade classifier. FPGA is reconfigurable, cost efficient and more flexible than other hardware platforms. The system is implemented using model-based design (Simulink) and hardware software co-design. Proposed design is analyzed on a Xilinx board.

Traffic signs detection and segmentation is one of the important parts of advanced driving assistance system. But there are predictable difficulties in detecting traffic signs from images or videos from car cameras owing to the next reasons: traffic signs are usually small-sized or medium-sized objects, and there is quantity imbalance between different traffic signs in the existed public data sets. Therefore, two main developments have been proposed in this paper. Firstly, an improved TT-100K-HHU traffic sign data set based on TT-100K is constructed. New images are collected from the Tencent Street View and labeled by Labelme software. Secondly, an improved Mask R-CNN is presented by revising the structure. More specific, feature pyramid network (FPN) is introduced into the backbone network of Mask R-CNN to achieve the fusion of feature maps at multiple scales, which can improve the representation abilities of network for objects with small or medium size. And in the prediction network, multiple cascaded Box Heads are applied to acquire more accurate location predictions and segmentation results. Experimental results show that the performance of the improved Mask R-CNN network is better than the existing algorithms.

This paper presents a deep convolutional neural network architecture to classify the traffic signs of the GTSRB dataset. Our method uses a very light architecture with a few number of parameters that achieve good results without the need of hard computation. To get at our goal, we use a filter bank. The aim of which being to extract more features, which will be used as input to a fully connected classifier. The recognition rate of our model gets an accuracy of 99.15%, overpassing the human performance being 98.81%. This way, LiteNet competes the best state of art architectures since our approach uses less memory and less computation.

Traffic sign detection is crucial for enabling autonomous vehicles to navigate in real-world streets, which must be carried out with high accuracy and in real-time. CNNs have become one of the standard approaches for traffic sign detection research in recent years. The use of CNNs has allowed the development of traffic sign detectors that are capable of achieving prediction accuracies similar to those of human drivers. However, most CNNs do not run in real-time due to the high number of computational operations involved during the inference phase. This hinders the deployment of CNNs in autonomous vehicles despite their high prediction accuracy. In this paper, we explore BNNs to tackle this problem. BNNs binarize the full-precision weights and activations of a CNN, drastically reducing the complexity of the computational operations required for inference, while at the same time maintaining the architectural parameters, as well as spatial dimensions of the input image. This reduces the memory required to run the model and enables faster inference time. We carry out in-depth studies on applying BNNs for traffic sign detection using real-world datasets. We observe an improvement of 11.63 × for normalized compute complexity, while suffering only 3.93 pp in detection accuracy on GTSDB dataset.

Autonomna vožnja jedno je od zanimljivijih istraživačkih područja modernog doba, a prepoznavanje prometnih znakova jedno je od važnijih i krucijalnih problema u tom istraživanju. U okviru ovog rada istraženi su algoritmi YOLOv3 i ResNet50 koji surađuju u procesu lokalizacije i klasifikacije prometnih znakova u slikama sa stvarnim uvjetima na cestama. YOLOv3 izvršava lokalizaciju prometnih znakova dok ResNet50 te znakove klasificira u jenu od 43 klase prometnih znakova. Za učenje YOLOv3 i ResNet50 korišteni su skupovi podataka GTSDB i GTSRB koji sadržavaju njemačke prometne znakove. Također, prikazani su rezultati rada naučenih modela. Autonomous driving is one of the most interesting research areas of modern age and traffic signs detection is one of the most important and crucial problem in that research. In this paper, the algorithms YOLOv3 and ResNet50, that cooperate in the process of localization and classification of traffic signs in images with real road conditions, were researched. YOLOv3 performs the traffic signs localization while ResNet50 classifies those signs into one of 43 traffic signs classes. The GTSDB AND GTSRB data sets containing German traffic signs were used to train YOLOv3 and ResNet50. Also, the work results of the trained models were presented.

Traffic sign detection based on image and video data is critical, which captures real-time traffic road information for autonomous vehicle. With the rapid development of CNN, more and more CNN-based detectors have promoted general object detection. However, these mainstream detectors still suffer from small object detection task because of small size and fuzzy representation. Traffic signs are representative small object on road scenes causing a rigid challenge for autonomous driving perception system. In this paper, traffic sign detection (TSD) is regard as a small object detection task. We propose a feature fusion method via cross-connection to enhance feature representation. In addition, contextual information searched by dilated convolution is also used to support small traffic sign detection. We have implemented our modules into Faster R-CNN and evaluated effectiveness of proposed method on Tsinghua-Tencent 100K dataset. Our experimental results prove that the feature fusion method via cross connection and contextual information improve detection result of small traffic sign.

As a component of the traffic control plan, traffic signs on highways offer drivers necessary information. Unfortunately, many signs are unfamiliar to or misunderstood by drivers, especially when lacking a setting method, this includes exit advance guide signs in tunnels. These are generally set in roadbed sections, but space limitations in tunnels dictate that they must be set differently. To evaluate the effect of the setting method, an experiment was designed and conducted, during which the eye movements of 44 drivers with different familiarity levels were tracked. Twenty-two of the drivers had not previously participated in any experiment involving exit advance guide signs in highway tunnels, while 22 of them had. Time period data were analyzed, including data from before the sign appeared, when it appeared, and when it disappeared. Based on area division and Markov theory, attributes related to gaze transition were obtained, including one- and two-step gaze transition probabilities and area gaze probabilities. The results showed that gaze transition was confirmed to be significantly different between the three periods and between the drivers. Features extracted from eye movement characteristics, gaze transition paths, and gaze areas demonstrated that visual attention is more dispersed in familiar drivers during the lane-change intention period. Therefore, signs should be placed on the left wall of the highway tunnel.

In order to recognize prohibition traffic signs, based on the analysis of the color occupancy of prohibition traffic signs, this paper proposes a method to recognize the prohibition traffic signs based on the feature of Histogram of Oriented Gradient (HOG) and Support Vector Machine (SVM). The recognition method is mainly divided into three steps: the first step is image preprocessing, which realizes the size normalization processing, grayscale processing and Gamma correction of the image; the second step is the feature extraction of HOG; the third step is the recognition of prohibition traffic signs based on SVM. In the design and implementation of the prohibition traffic sign classifier, the prohibition traffic sign image training after linear transformation is used to train 42 binary classifiers, and then based on these 42 classifiers, the prohibition traffic sign classifier is constructed and implemented. Finally, the self-built data set was used to test and analyze the prohibition traffic sign recognition method, and the overall recognition accuracy rate was 90.2%.