Dissertations / Theses on the topic 'Detection of road lane lines'

The objective of this dissertation is to develop a Multi-Camera Lane Departure Warning (MCLDW) system and a framework to evaluate it. A Lane Departure Warning (LDW) system is a safety feature that is included in a few luxury automobiles. Using a single camera, it performs the task of informing the driver if a lane change is imminent. The core component of an LDW system is a lane detector, whose objective is to find lane markers on the road. Therefore, we start this dissertation by explaining the requirements of an ideal lane detector, and then present several algorithmic implementations that meet these requirements. After selecting the best implementation, we present the MCLDW methodology. Using a multi-camera setup, MCLDW system combines the detected lane marker information from each camera's view to estimate the immediate distance between the vehicle and the lane marker, and signals a warning if this distance is under a certain threshold. Next, we introduce a procedure to create ground truth and a database of videos which serve as the framework for evaluation. Ground truth is created using an efficient procedure called Time-Slicing that allows the user to quickly annotate the true locations of the lane markers in each frame of the videos. Subsequently, we describe the details of a database of driving videos that has been put together to help establish a benchmark for evaluating existing lane detectors and LDW systems. Finally, we conclude the dissertation by citing the contributions of the research and discussing the avenues for future work.

O uso de sistemas de assistência ao motorista (DAS) baseados em visão tem contribuído consideravelmente na redução de acidentes e consequentemente no auxílio de uma melhor condução. Estes sistemas utilizam basicamente uma câmera de vídeo embarcada (normalmente fixada no para-brisa) com o propósito de extrair informações acerca da rodovia e ajudar o condutor num melhor processo de dirigibilidade. Pequenas distrações ou a perda de concentração podem ser suficientes para que um acidente ocorra. Este trabalho apresenta uma proposta para o desenvolvimento de algoritmos para extrair informações sobre a sinalização em rodovias. Mais precisamente, serão abordados algoritmos de calibração de câmera explorando a geometria da pista, de extração da marcação de pintura (sinalização horizontal) e detecção e identificação de placas de trânsito (sinalização vertical). Os resultados experimentais indicam que o método de calibração de câmera alcançou bons resultados na obtenção dos parâmetros extrínsecos com erros inferiores a 0:5 . O erro médio encontrado nos experimentos com relação a estimativa da altura da câmera foi em torno de 12 cm (erro relativo aproximado de 10%), permitindo explorar o uso da realidade aumentada como uma possível aplicação. A acurácia global para a detecção e reconhecimento da sinalização horizontal (marcas seccionadas, contínuas e mistas) foi acima de 96% perante uma diversidade de situações apresentadas, tais como: sombras, variação de iluminação, degradação do asfalto e pintura. O uso da câmera calibrada para a detecção da sinalização vertical contribui para delimitar o espaço de varredura da janela deslizante do detector, bem como realizar a procura por placas em uma única escala para cada região de busca, caracterizada pela distância ao veículo. Os resultados apresentados reportam uma taxa global de classificação de aproximadamente 99% para o sinal de proibido ultrapassar, considerando-se uma base de dados limitada a 962 amostras.
The use of driver assistance systems (DAS) based on computer vision has helped considerably in reducing accidents and consequently aid in better driving. These systems primarily use an embedded video camera (usually fixed on the windshield) for the purpose of extracting information about the highway and assisting the driver in a better handling process. Small distractions or loss of concentration may be sufficient for an accident to occur. This work presents the development of algorithms to extract information about traffic signs on highways. More specifically, this work will tackle a camera calibration algorithm that exploits the geometry of the road track, algorithms for the extraction of road marking paint (lane markings) and detection and identification of vertical traffic signs. Experimental results indicate that the proposed method for obtaining the extrinsic parameters achieve good results with errors of less than 0:5 . The average error in our experiments, related to the camera height, were around 12 cm (relative error around 10%). Global accuracy for the detection and classification of road lane markings (dashed, solid, dashed-solid, solid-dashed or double solid) were over 96%. Finally, our camera calibration algorithm was used to reduce the search region and to define the scale of a slidingwindow detector for vertical traffic signs. The use of the calibrated camera for the detection of traffic signs contributes to define the scanning area of the sliding window and perform a search for signs on a unique scale for each region of interest, determined by the distance to the vehicle. The results reported a global classification rate of approximately 99% for the no overtaking sign, considering a limited of 962 samples.

The interest for autonomous driving assistance, and in the end, self-driving cars, has increased vastly over the last decade. Automotive safety continues to be a priority for manufacturers, politicians and people alike. Visual-based systems aiding the drivers have lately been boosted by advances in computer vision and machine learning. In this thesis, we evaluate the concept of an end-to-end machine learning solution for detecting and classifying road lane markings, and compare it to a more classical semantic segmentation solution. The analysis is based on the frame-by-frame scenario, and shows that our proposed end-to-end system has clear advantages when it comes detecting the existence of lanes and producing a consistent, lane-like output, especially in adverse conditions such as weak lane markings. Our proposed method allows the system to predict its own confidence, thereby allowing the system to suppress its own output when it is not deemed safe enough. The thesis finishes with proposed future work needed to achieve optimal performance and create a system ready for deployment in an active safety product.

One of the main challenges for advanced driver assistance systems (ADAS)is the environment perception problem. One factor that makes ADAS hardto implement is the large amount of different conditions that have to betaken care of. The main sources for condition diversity are lane and roadappearance, image clarity issues and poor visibility conditions. A review ofcurrent lane detection algorithms has been carried out and based on that alane detection algorithm has been developed and implemented on a mixedcriticality platform. The thesis is part of a larger group project consisting offive master thesis students creating a demonstrator for autonomous platoondriving. The final lane detection algorithms consists of preprocessing stepswhere the image is converted to gray scale and everything except the regionof interest (ROI) is cut away. OpenCV, a library for image processing hasbeen utilized for edge detection and hough transform. An algorithm for errorcalculations is developed which compares the center and direction of the lanewith the actual vehicle position and direction during real experiments. Thelane detection system is implemented on a Raspberry Pi which communicateswith a mixed criticality platform through UART. The demonstrator vehiclecan achieve a measured speed of 3.5 m/s with reliable lane keeping using thedeveloped algorithm. It seems that the bottleneck is the lateral control ofthe vehicle rather than lane detection, further work should focus on controlof the vehicle and possibly extending the ROI to detect curves in an earlierstage.
En stor utmaning för avancerade förarstödsystem (ADAS) är problemet med uppfattning av miljön runt omkring. En faktor som gör ADAS svårt att implementera är den stora mängd olika förhållanden som måste tas hand om. De största källorna till olikheter är utseendet på körfältet och vägen, dåliga siktförhållanden samt otydliga bilder. En granskning av nuvarande algoritmer för körfältsdetektering har utförts och baserat på den har en körfältsdetekteringsalgoritm utvecklats och implementerats på ett blandkritiskt system. Avhandlingen är en del av ett större grupprojekt bestående av fem mastersstudenter som ska skapa en demonstrator för autonom konvojkörning. Den slutgiltiga körfältsdetekteringsalgoritmen består av förbehandlingssteg, där bilden konverteras till gråskala och allt utom intresseområdet är bortklippt. OpenCV, ett bibliotek för bildbehandling har använts för kantdetektering och houghtransformation. En algoritm som jämför körfältets mittpunkt och riktning med fordonets faktiska position och riktning har utvecklats och används i experiment för kontroll av fordonet. Körfältsdetekteringsalgoritmen har implementeras på en Raspberry Pi som kommunicerar med en blandkritisk plattform genom UART. Demo-fordonet kan uppnå en uppmätt hastighet på 3,5 m/s med pålitlig väghållning med den utvecklade algoritmen. Det verkar som att flaskhalsen är kontroll av fordonet i sidled och inte körfältsdetektering, ytterligare arbete bör fokusera på kontroll av fordonet och eventuellt utöka synfältet för att detektera kurvor i ett tidigare skede.

A self-driven car is a vehicle that can drive without human intervention by making correct decisions based on the environmental conditions. Since the innovation is in its beginning periods, totally moving beyond the human inclusion is still a long shot. However, rapid technological advancements are being made towards the safety of the driver and the passengers. One such safety feature is a Lane Detection System that empowers vehicle to detect road lane lines in various climate conditions. This research provides a feasible and economical solution to detect the road lane lines while driving in a sunny, rainy, or snowy weather condition. An algorithm is designed to perform real time road lane line detection on a low voltage computer that can be easily powered in a regular auto vehicle. The algorithm runs on a RaspberryPi computer placed inside the car. A camera, attached to the vehicle’s windshield, captures the real time images and passes them to the RaspberryPi for processing. The algorithm processes each frame and determines the lane lines. The detected lane lines can be viewed on a 7 inch display screen connected to the Raspberry Pi. The entire system is mounted inside a Jeep Wrangler to conduct the experiments and is powered by the vehicle’s standard charger of 12V-15V power supply. The algorithm provides approximately 97% accurate detection of road lane lines in all weather conditions.

With the rapid improvement of the Advanced Driver Assistant System (ADAS), autonomous driving has become one of the most common hot topics in recent years. While driving, many technologies related to autonomous driving choose to use the sensors installed on the vehicle to collect the information of road status and the environment outside. This aims to warn the driver to perceive the potential danger in the fastest time, which has become the focus of autonomous driving in recent years. Although autonomous driving brings plenty of conveniences to people, the safety of it is still facing difficulties. During driving, even the experienced driver can not guarantee focus on the status of the road all the time. Thus, lane departure warning system (LDWS) becomes developed. The purpose of LDWS is to determine whether the vehicle is in the safe driving area. If the vehicle is out of this area, LDWS will detect it and alert the driver by the sensors, such as sound and vibration, in order to make the driver back to the safe driving area. This thesis proposes a novel lightweight LDWS model LEHA, which divides the entire LDWS into three stages: image preprocessing, lane detection, and lane departure recognition. Different from the deep learning methods of LDWS, our LDWS model LEHA can achieve high accuracy and efficiency by relying only on simple hardware. The image preprocessing stage aims to process the original road image to remove the noise which is irrelevant to the detection result. In this stage, we apply a novel algorithm of grayscale preprocessing to convert the road image to a grayscale image, which removes the color of it. Then, we design a binarization method to greatly extract the lane lines from the background. A newly-designed image smoothing is added to this stage to reduce most of the noise, which improves the accuracy of the following lane detection stage. After obtaining the processed image, the lane detection stage is applied to detect and mark the lane lines. We use region of interest (ROI) to remove the irrelevant parts of the road image to reduce the detection time. After that, we introduce the Canny edge detection method, which aims to extract the edges of the lane lines. The last step of LDWS in the lane detection stage is a novel Hough transform method, the purpose of it is to detect the position of the lane and mark it. Finally, the lane departure recognition stage is used to calculate the deviation distance between the vehicle and the centerline of the lane to determine whether the warning needs to turn on. In the last part of this paper, we present the experiment results which show the comparison results of different lane conditions. We do the statistic of the proposed LDWS accuracy in terms of detection and departure. The detection rate of our proposed LDWS is 98.2% and the departure rate of it is 99.1%. The average processing time of our proposed LDWS is 20.01 x 10⁻³s per image.

This thesis describes the design and subsequent implementation of a unit for traffic monitoring using Raspberry PI. First section provides a quick overview of assistance systems, which use a road lane detection techniques. Next there is a description of two diferent methods for road lane detection. Follow the description of monitoring scene. Then the work describe the practical part including the design and realization of supporting electronics, selecting of each components, including the modifying of cameras, mechanical design and creating of unit. Another section is about selection and installation of appropriate software components necessary for running of the unit and the selection of development tools for creating user application. After description of graphical user interafce, there is a description of road lanes detection algorithm. At the end of the thesis is summarized a reliability of unit in real traffic situation. At the appendix there are technical drawings, describing the unit.

碩士
國立交通大學
電控工程研究所
107
In this thesis we consider one of challenge cases in a lane detection system, which is channelizing line, and propose an efficient method to not only keep the pixels belonging to the lane located on the channelizing line, but also detect and model the lanes which were filtered from the observation. Channelizing lines are used to guide drivers where travel in the same direction for permitting on both sides, such as entrance and exit ramps. Channelizing line usually has same intensity with the lane, but has pavement markings, solid white lines with wide diagonal lines or chevrons within two channelizing lines, which are usually influence the detection result in several previous work. The experiments show that the proposal has a good performance even passing through several channelizing lines. While passing through single channelizing line zone, the average accuracy can reach to 91.11%. Experiments show that our approach reaches competitive performances on channelizing lines.

碩士
銘傳大學
資訊傳播工程學系碩士班
103
Lane line detection and tracking plays an important role in the automatic vehicle monitoring. However, the detection for non-linear lanes may lead to significant detection error. This thesis studies an adaptive road mask generation method to overcome changes of curvature of land road. Furthermore, the adaptive road mask is generated by applying the fast vanishing-point detection scheme. This method is highly computationally efficient, robust to lane change, and adaptive to the captured road scenario. Experimental results with a variety of curved lanes videos confirmed that this method can get good detection results.

碩士
國立宜蘭大學
電子工程學系碩士班
106
According to studies, on traffic roads the accident rate at night was higher than during day. In particular, human error is still the major cause of accidents and incidents. In order to solve the accidents caused by driving, the application of Advanced Driver Assistance Systems (ADAS) are becoming more important than ever. ADAS developments for vision systems are a key focus that detect vehicles on the road in front of one's own by using computer vision technologies. There are designed to warn of possible dangers and prevent traffic accidents. This paper is focused on taillight to develop a computer vision-based nighttime vehicle detection. We present a novel adaptive threshold technique based on fuzzy system and a dynamic threshold is calculated using Otsu's method. In order to ensure correctness of detecting, we also create a new detection model to determine the position of target vehicles in road region. We propose a method that can help to overcome problems such as the temporary loss of the target position caused by using traditional detection model. In testing phase, we propose a dynamic threshold and compare two other studies through different traffic scenarios for simulation. Experiment results indicate that, the proposed scheme which can significantly improves vehicle detection rates under complex illumination changes. It achieves a good result to extract foreground objects from sample nighttime traffic scenes. We also explore how to deformable detection model to detect vehicles. These method can be used to effectively reduce the miss rate of our vehicle detection.

碩士
國立臺灣科技大學
電子工程系
107
This thesis investigates the problem of detecting lane and road markings for self-driving cars. We pose this problem as a semantic segmentation problem, which is addressed by the deep learning scheme. Self-driving cars must analyze road scene in real-time. Therefore, our research focuses on the new neural network models for semantic segmentation, which can reduce the inference time while maintaining the detection accuracy. Inspired by Enet and Bisnet, we propose two neural network models for this problem. First, we design a new smaller model by removing several redundant bottleneck layers from Enet and then applying the knowledge distillation scheme, which uses the results of the original complex and high- precision E -net as a soft tag to guide the training of the new small, and fast neural network model to boost its detection accuracy. Next, we propose a novel network architecture with two parallel networks: one network can infer the spatial features of the inputs, while the other network is responsible for extending the reception fields and encoding the context messages, and then combining the results from both networks to produce the final output. To assess the performance of the two proposed network models, we perform experiments on the ITRI road marking datasets , the experimental results demonstrate the superiority of our proposed network models even using a smaller number of training datasets.

碩士
國立交通大學
電子研究所
106
Lane mark detection is an essential component in the road scene analysis for Advanced Driver Assistant System (ADAS). Although the deep-learning based road scene segmentation can achieve very high accuracy, limited by the onboard computing power, it is still a challenge to reduce system complexity and to maintain high accuracy at the same time. In this thesis, we incorporate the deep convolutional neural network into a lane detection algorithm in order to extract the robust lane mark features. To improve its performance with a target of lower complexity, we investigate the advantages and disadvantages of several popular CNN architectures in terms of speed and storage. We start from SegNet with VGG and continue to study the Fully Convolutional Network (FCN), ResNet, and DenseNet. Through detailed experiments, we pick up favorable components from the existing architectures and at the end, a light network architecture is thus constructed based on the structure of DenseNet. Our proposed network demonstrates a real-time testing (inferencing) ability and it maintains an accuracy comparable with most previous systems. We test our system on several datasets including the challenging Cityscapes dataset (resolution of 1024×512) with a mIoU of about 69.1 %. We also design a more accurate model but at the price of a slower speed with a mIoU of about 72.9 % on the CamVid dataset. Moreover, we also design a post-processing algorithm to group segments of a lane into a connected curve and construct a 3rd- order polynomial model to fit into a curved lane. Our system shows promising results on the captured road scenes.

In this modern era, land transports are increasing dramatically. Moreover, self-driven car or the Advanced Driving Assistance System (ADAS) is now the public demand. For these types of cars, road conditions detection is mandatory. On the other hand, compared to the number of vehicles, to increase the number of roads is not possible. Software is the only alternative solution. Road Conditions Detection system will help to solve the issues. For solving this problem, Image processing, and machine learning have been applied to develop a project namely, Detection of Road Conditions Using Image Processing and Machine Learning Techniques for Situation Awareness. Many issues could be considered for road conditions but the main focus will be on the detection of potholes, Maintenance sings and lane. Image processing and machine learning have been combined for our system for detecting in real-time. Machine learning has been applied to maintains signs detection. Image processing has been applied for detecting lanes and potholes. The detection system will provide a lane mark with colored lines, the pothole will be a marker with a red rectangular box and for a road Maintenance sign, the system will also provide information of aintenance sign as maintenance sing is detected. By observing all these scenarios, the driver will realize the road condition. On the other hand situation awareness is the ability to perceive information from it’s surrounding, takes decisions based on perceived information and it makes decision based on prediction.

Road detection is a crucial concern in Autonomous Navigation and Driving Assistance. Despite the multiple existing algorithms to detect the road, the literature does not offer a single effective algorithm for all situations. A global more robust set-up would count on multiple distinct algorithms running in parallel, or even from multiple cameras. Then, all these algorithms’ outputs should be merged or combined to produce a more robust and informed detection of the road lane, so that it works in more situations than each algorithm by itself. This dissertation integrated in the ATLAS-CAR2 project, developed at the University of Aveiro, proposes a ROS-based architecture to manage and combine multiple sources of lane detection algorithms ranging from the algorithms that return the spatial localization of the road lane lines and those whose results are the navigable zone represented as a polygon. The architecture is fully scalable and has proved to be a valuable tool to test and parametrise individual algorithms. The combination of the algorithms’ results used in this work uses a confidence based merging of individual detections.
A deteção de estradas é uma questão crucial na Navegação Autónoma e na Assistência à Condução. Apesar de os múltiplos algoritmos existentes para detetar a estrada, a literatura não oferece um único algoritmo eficaz para todas as situações. Uma configuração global mais robusta incorporaria vários algoritmos distintos e executados em paralelo, ou mesmo baseado em múltiplas câmaras. Então, todos os resultados destes algoritmos devem ser fundidos ou combinados para produzir uma deteção mais robusta e informada da via da estrada, para que funcione em mais situações do que cada algoritmo funcionando individualmente. Esta dissertação integrada no projeto ATLASCAR2, desenvolvido na Universidade de Aveiro, propõe uma arquitetura baseada em ROS para gerir e combinar múltiplas fontes de algoritmos de deteção de vias da estrada, desde algoritmos que devolvem a localização espacial da faixa de rodagem até àqueles cujos resultados são a zona navegável representada como um polı́gono. A arquitetura é totalmente escalável e provou ser uma ferramenta valiosa para testar e parametrizar algoritmos individuais. A combinação dos resultados dos algoritmos utilizados neste trabalho utiliza uma combinação de deteções individuais baseada na confiança.
Mestrado em Engenharia Mecânica