Road Lane Research Articles

심층학습 기반 차선 분할과 최적 추정기를 이용한 자율주행차 차선인식 기법

Road lane-line detection systems are crucial for assisting human drivers in various driving scenes. Nowadays, these systems have become more critical for autonomous cars. In recent years, many advanced lane detection and tracking methods have been studied. However, most approaches focus on detecting lanes by performing conventional image processing on a single frame. Often, the lack of distinguishing features and the presence of extreme weather conditions such as dense rain or snow makes these algorithms lose their accuracy. In this work, deep learning is merged as a tool to address these issues. A modified version of the LaneNet algorithm was designed as a road lane detector, in which the front image was passed as input to the encoder model. Consequently, the road lanes could be robustly detected even if they were discrete or unclear. For lane curvature estimation, a Kalman filter was utilized, which took the results obtained from the LaneNet detection module as input. The detection results were compared with those obtained using the conventional Hough transformation. The obtained results demonstrate the effectiveness of our approach.

Self-Driving Car

Abstract: Every person in this world is concerned about being safe. Increasing safety and reducing road accidents, thereby saving lives are of great interest in the context of Advanced Driver Assistance Systems. Among the complex and challenging tasks of future road vehicles is road lane detection or road boundaries detection. In driving assistance systems, obstacle detection especially for moving object detection is a key component of collision avoidance. Many sensors can be used for obstacle detection and lane detection, such as laser, radar and vision sensors. The most frequently used principal approach to detect road boundaries and lanes using a vision system on the vehicle. Detecting all kinds of obstacles on the road, mainly include the IPM (Inverse Perspective Mapping) method. The system acquires the front view using a camera mounted on the vehicle then applying a few processes in order to detect the lanes and objects. A versatile methodology is used in order to detect the lanes and objects. In our research we have developed a simple heuristic method which is more robust in both lane detection object detection and tracking in video. In this method we use clustering methodology to group the detected points in case of lane detection. Heuristic gives effective results in detection and tracking of multiple vehicles at a time irrespective to the distance.

A real-time obstacle avoidance and path tracking strategy for a mobile robot using machine-learning and vision-based approach

In this paper, an obstacle avoidance and target tracking method for both indoor and outdoor mobile robots in dynamic environment is presented, and it aims to enhance autonomous navigation capability of such robots. In the proposed method, image processing and machine-learning approaches are considered. Since obstacles have differences in color and texture and in order to identify non-navigable areas, a monocular onboard camera is used to capture the road lanes by dealing with an image processing technique. The position of the robot with respect to road lane center during navigation is calculated on the basis of a proposed fuzzy logic rules set. In order to provide fast and robust computation, the Haar cascade classifier-based machine-learning technique has been exploited to detect the different sizes and shapes of the obstacles faced by the robot during its movement from source to destination. The dynamic model of a four-wheel mobile robot is initially developed using bond graph theory and then, the proposed obstacle-avoidance strategy is applied. The effectiveness and performance of the proposed method are tested under various simulation and experimentation scenarios. The behavior of the mobile robot for detecting and avoiding static obstacle for single and double road lane change environment is analyzed and discussed.

Sharpening and Detection of Road Studs for Intelligent Vehicles at Nighttime

Intelligent vehicles need to survive against various environments on roads such as sunlit, unclear, showery, shadowy, and inside tunnel conditions. This research designs a robust approach for detecting road studs at nighttime, which is the combination of various statistical methods. This detection approach is a unique approach developed for the detection of road lanes instead of road-painted lanes. Therefore, we detect road studs (cat eyes) instead of the painted lanes on roads at nighttime as the road studs have higher intensities at nighttime. First, we utilized Butterworth low-pass filter in order to sharpen the images. Second, we converted the image to grayscale and extracted the corresponding region of interest (ROI) from it. Then, the Canny edge detection algorithm was applied to create boundary lines in images. Finally, the Hough transform was applied to detect the desired lanes in the images, which are the road studs, and hence we successfully detected the road studs in images. We have used our own dataset for the stud’s detection, which considered most of the limitations of the previous datasets. Also, the dataset was collected in naturalistic environments at nighttime. The experimental result presents that the designed approach is accurate and robust for road stud detection against nighttime.

Deep embedded hybrid CNN–LSTM network for lane detection on NVIDIA Jetson Xavier NX

In recent years, lane detection has become one of the most important factors in the progress of intelligent vehicles. To deal with the challenging problem of low detection precision and real-time performance of most traditional systems, we proposed a real-time deep lane detection system based on CNN Encoder–Decoder and Long Short-Term Memory (LSTM) networks for dynamic environments and complex road conditions. The CNN Encoder network is used to extract deep features from a dataset and to reduce their dimensionality. A corresponding decoder network is used to map the low resolution encoder feature maps to dense feature maps that correspond to road lane. The LSTM network processes historical data to improve the detection rate through the removal of the influence of false alarm patches on detection results. We propose three network architectures to predict the road lane: CNN Encoder–Decoder network, CNN Encoder–Decoder network with the application of Dropout layers and CNN Encoder–LSTM-Decoder network that are trained and tested on a public dataset comprising 12764 road images under different conditions. Experimental results show that the proposed hybrid CNN Encoder–LSTM-Decoder network that we have integrated into a Lane-Departure-Warning-System (LDWS) achieves high prediction performance namely an average accuracy of 96.36%, a Recall of 97.54%, and a F1-score of 97.42%. A NVIDIA Jetson Xavier NX supercomputer has been used, for its performance and efficiency, to realize an Embedded Deep LDWS.

Research on road vehicle and lane line detection based on improved YOLOv4-tiny algorithm

In order to improve the speed of road vehicle and lane line detection, a road vehicle and lane line detection method with improved YOLOv4-tiny algorithm is proposed by using K-means++ clustering algorithm instead of the original K-means algorithm. Using the Mosaic data enhancement method, four images are randomly deflated and then stitched into one image to enrich the detection target background. The optimal weight values are derived by multi-scale training using the GPU through image feature extraction of the BDD10K dataset to achieve the detection of vehicle and lane line targets in the images. The results show that the improved YOLOv4-tiny algorithm achieves a detection speed of 134 FPS and an average accuracy of 77.84% mAP in highway lane line detection. After comparison, the detection speed of the improved algorithm is significantly improved, effectively improving the efficiency of highway vehicle and lane line detection.

Air quality analysis on a frequently congested road section

This paper will focus on analysing the air quality on a frequently congested road section in the vicinity of Timisoara, Romania. The placement of the measurement equipment will be made next to a public transport station measuring the extent of different pollutants which affect the well-being of passengers waiting in these stations. The traffic congestion is present due to the unloading from the highway, reduced capacity of only a one lane road and because of the high amount of industrial area workplaces. The results presented will be part of the PhD thesis of the main author.

A Lane-Level Road Marking Map Using a Monocular Camera

The essential requirement for precise localization of a self-driving car is a lane-level map which includes road markings (RMs). Obviously, we can build the lane-level map by running a mobile mapping system (MMS) which is equipped with a high-end 3D LiDAR and a number of high-cost sensors. This approach, however, is highly expensive and ineffective since a single high-end MMS must visit every place for mapping. In this paper, a lane-level RM mapping system using a monocular camera is developed. The developed system can be considered as an alternative to expensive high-end MMS. The developed RM map includes the information of road lanes (RLs) and symbolic road markings (SRMs). First, to build a lane-level RM map, the RMs are segmented at pixel level through the deep learning network. The network is named RMNet. The segmented RMs are then gathered to build a lane-level RM map. Second, the lane-level map is improved through loop-closure detection and graph optimization. To train the RMNet and build a lane-level RM map, a new dataset named SeRM set is developed. The set is a large dataset for lane-level RM mapping and it includes a total of 25157 pixel-wise annotated images and 21000 position labeled images. Finally, the proposed lane-level map building method is applied to SeRM set and its validity is demonstrated through experimentation.

A Method for Automatic Lane Detection using a Deep Network

Lane detection is amongst the most important operations during the automatic driving process. This process aims to detect lane lines to control the vehicle’s direction in a specific lane on the road and it can be effective in preventing accidents. Besides being online, the most requirement for lane detection methods in automatic driving is their high accuracy. The use of deep learning to create fully automated systems in lane detection has been done extensively. Automated learning-based methods used for road lane detection are often of supervised type. One of the disadvantages of these methods, despite their excellent accuracy, is that they need a set of labeled data, which limits the development procedure of the lane detection system, and most importantly establishing a standard set of labeled data is very time-demanding. The recommended solution is to use appropriate learning approaches that can be used to achieve the relative accuracy of the supervised approaches and to improve their speed. It also enables us to use different datasets without a constraint label that the tagged dataset would create in the algorithm development relative to the new dataset. In this research, we present an automatic semi-supervised learning method using deep neural networks to extract the data of lane lines, using labeled (mask for detected lines) and unlabeled datasets. The results demonstrate the suitable accuracy of the adopted method according to the proposed approach and also improves its computational complexity due to the significant reduction in the number of teachable network parameters.

Computer Vision Based Road Lane Line Detection

Computer Vision Based Road Lane Line Detection

SUSTAINABLE TRANSPORTATION SYSTEM

Global transport policy and planning now prioritise sustainability above all else. Similar to sustainable development, the definition of sustainable transportation is wide, allowing for the designation of some kilometer and policies as “sustainable” even when they follow standard operating procedures. Therefore, it is imperative that sustainability concepts be properly integrated into and applied to the transportation sector. Frameworks for performance measurement provide an efficient means of doing this. Much work has gone into comprehending and implementing the idea of sustainable development in the transport sector during the last 20 years.1. We may get knowledge from a plethora of study and experience in this area. However, there hasn’t been much headway in implementing more environmentally friendly modes of transportation. Given the rising demand for infrastructure and mobility in many areas, the detrimental effects of transport are probably only going to get worse (Dulca 2013; AfDB et al. 2012). For instance, the expansion in passenger and freight traffic, particularly in emerging nations, is predicted to need over 25 million paved road lane kilometers and 335,000 rail track kilometers worldwide by 2050 (Dulca 2013). To put this into context, the total length of all road and railway networks worldwide would have increased by 60% as a result (ibid.). In line with these projections, oil consumption is anticipated to rise in 2035, mostly due to demand from China and India, while it is likely to decline in the Organisation for Economic Co-operation and Development.

New Ontology structure for intelligent controlling of traffic signals

This article proposes a novel ontology design for intelligent controlling of traffic signals, considering the investigated factors, crowded factors, road factors, visibility conditions, and emergency situations. Essentially, the proposed method uses video-based knowledge and key feature from a monocular video camera only, capturing footage from either a traffic signals perspective or the top of the road lane. The key factors and entities in the traffic scene are formed into an ontology, which has been evaluated using synthetic datasets to interpret challenging cases. Semantic features related to the key factors in the scene are obtained and fed to the ontology. The experimental results indicate that the proposed method is capable of controlling traffic signals more efficiently than the fixed intervals protocol.

Computer Vision Based Road Lane Line Detection

Computer Vision Based Road Lane Line Detection

An Enhanced System for Wrong-Way Driving Vehicle Detection with Road Boundary Detection Algorithm

Driving in the wrong direction is one of the main reasons that cause road accidents in Thailand. To efficiently detect wrong direction driving vehicles, we proposed a system that can track those moving vehicles from CCTVs using deep learning and image processing techniques. Our proposed system is composed of two main algorithms: Road Lane Boundary detection from CCTV algorithm (RLB-CCTV) and Majority-Based Correct Direction Detection algorithm (MBCDD). In this paper, we introduced the RLB-CCTV algorithm and the enhanced version of MBCDD algorithms for detecting road boundaries and wrong-direction driving vehicles. Based on our experiments, our system can identify the road lane boundary and track vehicles driving in the wrong direction on each lane with an accuracy of 96.61%.

An Economically Efficient Agent-based Autonomous Road Cleaner for Asian Cities

In the last couple of decades, autonomous road cleaners have been enormously attracting the world. For this purpose, numerous researchers have contributed to designing efficient, robust, and human-friendly cleaning mechanisms. However, their proposed approaches do not provide a cost-effective solution due to the deployment of exorbitant sensors and sophisticated infrastructure deployment. Hence, we propose to develop a cost-effective self-driving road that has the capability to clean the roads of major cities of Asia. The developed road cleaner efficiently detects road lanes using advanced machine vision techniques and estimates the angle to stay in the lane. Moreover, to control the speed of the road cleaner we have used a proportional–integral–derivative (PID) controller. The smart cleaning robot detects objects in its surroundings to prevent accidents in congested roads; thus it will have a function of anticollision and automatic avoidance. It also measures turning angle by performing computations on the position of road lanes and takes turns accordingly. To validate the efficiency of the proposed solution, we have performed a series of experiments considering different test cases.

Examining the effects of vertical outdoor built environment characteristics on indoor light pollution

Indoor light pollution threatens human health. This study examined how built environment characteristics affect nighttime light levels in the bedrooms of urban areas. Indoor illuminances were monitored in the bedrooms of 28 participants in the city of Daejeon, Korea, between 2 a.m. and 4:30 a.m. over five days. Five built environment variables surrounding the bedrooms under study were assessed. Results illustrate that residential floor area on floors of influence (±3 floors from the bedroom under study) increased indoor illuminance, but residential floor area on all floors, including those vertically distant from the bedroom, did not. Tree canopy percentage reduced indoor illuminance, suggesting that tree canopies block artificial light at night emitted at ground level. Non-residential floor area and length of road lanes did not have significant effects, further suggesting that tree canopies protect urban residents on upper floors from light directed upwards from streets and commercial facilities at ground level. These findings indicate that vertical separation of residential areas is an effective measure to mitigate light pollution, especially in high-density urban areas where horizontal separation is not feasible. These findings also imply that policy efforts should focus on reducing light trespass from neighbors at similar heights.

Ultrafine particle exposure for bicycle commutes in rush and non-rush hour traffic: A repeated measures study in Copenhagen, Denmark

Ultrafine particles (UFP), harmful to human health, are emitted at high levels from motorized traffic. Bicycle commuting is increasingly encouraged to reduce traffic emissions and increase physical activity, but higher breathing rates increase inhaled UFP concentrations while in traffic. We assessed exposure to UFP while cycling along a fixed 8.5 km inner-city route in Copenhagen, on weekdays over six weeks (from September to October 2020), during morning and afternoon rush-hour, as well as morning non-rush-hour, traffic time periods starting from 07:45, 15:45, and 09:45 h, respectively. Continuous measurements were made (each second) of particle number concentration (PNC) and location. PNC levels were summarized and compared across time periods. We used generalized additive models to adjust for meteorological factors, weekdays and trends. A total of 61 laps were completed, during 28 days (∼20 per time period). Overall mean PNC was 18,149 pt/cm3 (range 256–999,560 pt/cm3) with no significant difference between morning rush-hour (18003 pt/cm3), afternoon rush-hour (17560 pt/cm3) and late morning commute (17560 pt/cm3) [p = 0.85]. There was substantial spatial variation of UFP exposure along the route with highest PNC levels measured at traffic intersections (∼38,000-42000 pt/cm3), multiple lane roads (∼38,000-40000 pt/cm3) and construction sites (∼44,000-51000 pt/cm3), while lowest levels were measured at smaller streets, areas with open built environment (∼12,000 pt/cm3), as well as at a bus-only zone (∼15,000 pt/cm3). UFP exposure in inner-city Copenhagen did not differ substantially when bicycling in either rush-hour or non-rush-hour, or morning or afternoon, traffic time periods. UFP exposure varied substantially spatially, with highest concentrations around intersections, multiple lane roads, and construction sites. This suggests that exposure to UFP is not necessarily reduced by avoiding rush-hours, but by avoiding sources of pollution along the bicycling route.

Factors affecting severity of motorcycle accidents on Thailand's arterial roads: Multiple correspondence analysis and ordered logistics regression approaches

Thailand was classified as a middle-income country and ranked second highest in terms of road traffic fatality rate in the world in 2015. By 2018, this ranking went up to ninth in world which may be because of various earnest safety policies implementation, supporting road safety research and establishing a road safety directing center. However, crash fatality rate has considerably remained high until recent year, indicating a clear need for further related research. Considering severity of the crashes, the majority of fatal crashes involved the motorcycle road user. Therefore, motorcycle crashes are important issues and should be considered to mitigate fatality due to immoderate proportion of motorcycle road user and motorcyclist fatality. This study aims to identify factors that influence the severity of motorcycle accidents on Thailand's arterial roads by employing ordered logistic regression and multiple correspondence analysis. The results demonstrated that although both analyses were relatively different, they provided similar results. Age, road lanes, and helmet wearing were significant factors that influenced the severity of motorcycle accidents. The results could serve as reference for planning strategies or organizing campaigns to reduce and prevent death owing to road traffic accidents, which may enhance the overall image of road traffic safety in Thailand.

An efficient encode-decode deep learning network for lane markings instant segmentation

<span lang="EN-US">Nowadays, advanced driver assistance systems (ADAS) has been incorporated with a distinct type of progressive and essential features. One of the most preliminary and significant features of the ADAS is lane marking detection, which permits the vehicle to keep in a particular road lane itself. It has been detected by utilizing high-specialized, handcrafted features and distinct post-processing approaches lead to less accurate, less efficient, and high computational framework under different environmental conditions. Hence, this research proposed a simple encode-decode deep learning approach under distinguishing environmental effects like different daytime, multiple lanes, different traffic condition, good and medium weather conditions for detecting the lane markings more accurately and efficiently. The proposed model is emphasized on the simple encode-decode Seg-Net framework incorporated with VGG16 architecture that has been trained by using the inequity and cross-entropy losses to obtain more accurate instant segmentation result of lane markings. The framework has been trained and tested on a vast public dataset named Tusimple, which includes around 3.6K training and 2.7 k testing image frames of different environmental conditions. The model has noted the highest accuracy, 96.61%, F1 score 96.34%, precision 98.91%, and recall 93.89%. Also, it has <span>also obtained the lowest 3.125% false positive and 1.259% false-negative value, which transcended some of the previous researches. It is expected to</span> assist significantly in the field of lane markings detection applying deep neural networks.</span>

지역 행사로 인한 교통영향권역 설정에 관한 연구

Large-scale international events such as the Pyeongchang Winter Olympics and the Yeosu World Expo, as well as events and festivals held competitively by local governments, are held every year. Event planning and operation are important for the success of these events, but efforts to minimize the accessibility of visitors and the inconvenience of local residents around the event are essential. In particular, in downtown areas where there are many events or assemblies, when road lanes are blocked, road capacity decreases and traffic patterns in the area change, resulting in extreme traffic congestion. In this study, to examine how much road control caused by events or assemblies in downtown areas affects surrounding roads by using data, we propose a method of setting the influence area using MAPE (Mean Absolute Percentage Error). If the sphere of influence for each event can be identified in advance, each local government will be able to preemptively establish countermeasures for traffic congestion.