Efficient Lane Detection Research Articles

SGPLane: Efficient lane detection via sampled grid points for autonomous driving

Lane detection is one of the critical tasks for autonomous driving. Earlier works revolved around semantic segmentation and object detection with a special program for lanes. However, most methods still suffer from unstable post-processing algorithms which leads to a gap between camera input and downstream applications. In this paper, we propose a novel detection presentation form for lanes and design a simple network without any complicated post-process. Specifically, we use sampled gird points to express lane lines and construct a network for the special lane format, which is called SGPLane. Therefore, the network learns a regression branch and a confidence branch to realize end-to-end lane detection by setting the threshold confidence value. Our model is validated on the typical dataset and real-world driving scenes. Experiments on lane detection benchmarks show that our method outperforms previous methods with accuracy score of 96.84% on Tusimple dataset with high FPS and 76.85% on our real-world dataset.

HistWind2—An Algorithm for Efficient Lane Detection in Highway and Suburban Environments

HistWind2—An Algorithm for Efficient Lane Detection in Highway and Suburban Environments

Optimizations in Dynamic Origin Technique for Efficient Lane Detection for Autonomous Vehicles

Optimizations in Dynamic Origin Technique for Efficient Lane Detection for Autonomous Vehicles

A Novel Fast Responding Driver Assistance Technique with Efficient Lane Detection and Collision Avoidance Using Dynamic Feature Extraction in Any Environment

Road accidents caused by a driver's irresponsibility while driving is becoming increasingly common. Furthermore, if lanes are lacking on the road, even well-trained drivers may find it difficult to keep the lane while driving in low-light conditions. As a result, it is critical to design trustworthy, precise, and efficient mechanisms in the vehicle system that aid the driver in the event of a road collision. Almost every country in the world is attempting to conquer this greatest difficulty. The topic study article focuses on the crossroads and presents a more authentic and efficient diver aid system strategy in terms of lane departure alarm even when lanes are missing by taking prior lane patterns into account. In addition, research is being conducted in order to provide a speedy collision warning with nearly no false alarms. This study develops a fast-response spatial domain approach for detecting lanes on highways, and if lane markers are missing, virtual lanes are constructed using a novel suggested algorithm. In addition, for vehicle collision avoidance, the system estimates the distance, velocity, and direction from the frontal vehicle to itself. The suggested approach is evaluated on real-time videos in all environmental circumstances such as poor or bright sunlight, rain, and twilight, as well as on different road geometries such as straight and curving at various vehicle speeds. In every situation, the system has achieved more than 98.7 percent accuracy. Even if there are no markings on road, the system provides more accurate and reliable experimental results. Finally, the results are compared to several existing algorithms based on accuracy, precision, recall, and F1 score, along with processing time.

Efficient Lane Detection Technique Based on Lightweight Attention Deep Neural Network

For self-driving vehicles, detecting lane lines in changeable scenarios is a fundamental yet challenging task. The rise of deep learning in recent years has contributed to the thriving of autonomous driving. However, existing methods of lane detection based on deep learning have high requirements on computing environment, so their applicability is further restricted. This paper proposed an improved attention deep neural network (DNN), a lightweight semantic segmentation architecture catering for efficient computation in low memory, which contains two branches worked in different resolution. The proposed network integrates fine details captured by local interaction of pixels at high resolution into global contexts at low resolution, computing dense feature maps for prediction task. Based on the attributes of disparate feature resolution characteristics, different attention mechanisms are adopted to guide the network to effectively exploit the model parameters. The proposed network achieves comparable results with state-of-the-art methods on two popular lane detection benchmarks (TuSimple and CULane), with faster calculation efficiency at 259 frames-per-second (FPS) on CULane dataset, and the total number of model parameters only requires 1.57 M. This study provides a practical and meaningful reference for the application of lane detection in memory constrained devices.

An improved efficient model for structure-aware lane detection of unmanned vehicles

Lane detection algorithms require extremely low computational costs as an important part of autonomous driving. Due to heavy backbone networks, algorithms based on pixel-wise segmentation is struggling to handle the problem of runtime consumption in the recognition of lanes. In this paper, a novel and practical methodology based on lightweight Segmentation Network is proposed, which aims to achieve accurate and efficient lane detection. Different with traditional convolutional layers, the proposed Shadow module can reduce the computational cost of the backbone network by performing linear transformations on intrinsic feature maps. Thus a lightweight backbone network Shadow-VGG-16 is built. After that, a tailored pyramid parsing module is introduced to collect different sub-domain features, which is composed of both a strip pool module based on Pyramid Scene Parsing Network (PSPNet) and a convolution attention module. Finally, a lane structural loss is proposed to explicitly model the lane structure and reduce the influence of noise, so that the pixels can fit the lane better. Extensive experimental results demonstrate that the performance of our method is significantly better than the state-of-the-art (SOTA) algorithms such as Pointlanenet and Line-CNN et al. 95.28% and 90.06% accuracy and 62.5 frames per second (fps) inference speed can be achieved on the CULane and Tusimple test dataset. Compared with the latest ERFNet, Line-CNN, SAD, F1 scores have respectively increased by 3.51%, 2.84%, and 3.82%. Meanwhile, the result from our dataset exceeds the top performances of the other by 8.6% with an 87.09 F1 score, which demonstrates the superiority of our method.

A Review on the Artificial Bee Colony & Clahe Based Hybrid Hough Transform For Efficient Lane Detection

This paper represents that lane coloration has become popular in real time vehicular ad-hoc system. Lane detection is normally helpful to localize path limits, determine undesired lane variations, and to enable approximation of the upcoming geometry of the road. There are different types of methods that are used for detecting lines, curves and ellipses i.e. hough transform, clustering and curve fitting. The overall objective of this paper is to providing better performance in lane coloration algorithm. The methods developed till now are working effectively as well as giving good results. Keywords: Lane Detection, Hough Transform, Clustering, Curve Fitting

Efficient Lane Detection Algorithm using Different Filtering Techniques

Today, one of the largest areas of research and development in the automobile industry is road safety. Many deaths and injuries occur every year on public roads from accidents. However, the most dramatic fact is that, nearly all of the accidents are caused by driver mistakes. The main goal of the lane detection system is to reduce the number of these accidents. Remarkable amount of the current researches in this field focus on building autonomous driving systems. This research work presents an approach for improving the performance of lane detection algorithm by using different filtering techniques. This paper deals with an efficient ways of noise reduction in the images by using different filtering techniques. The main objective is to design, develop, implement and subsequently simulate an efficient lane detection algorithm which will provide high quality results in the case when noise is present in the signal. The scope of the dissertation is to implement lane detection algorithm without using any filter, to implement lane detection algorithm using median, wiener, and hybrid median filters. And to compare the performance on the basis of accuracy, specificity, BER, PSNR, without and with filters (Median, Wiener, Hybrid median filters). By giving some selected road images, experiments will be taken, that will be useful for performance comparison. A variety of tests will be performed using improved algorithm to test various aspects of the road images. Comparisons will be drawn among proposes strategy with well-known existing algorithms.

Environment classification and hierarchical lane detection for structured and unstructured roads

This study presents a hierarchical lane detection system with the ability to deal with both structured and unstructured roads. The proposed system classifies the environment first before applying suitable algorithms for different types of roads. Instead of dealing with all situations with one complicated algorithm, this hierarchical architecture makes it possible to achieve high accuracy with relatively simple and efficient lane detection algorithms. For environment classification, pixels with lane-marking colours are extracted as feature points. Eigenvalue decomposition regularised discriminant analysis is utilised in model selection and maximum likelihood estimation of Gaussian parameters in high-dimensional feature space. For structured roads, the extracted feature points are reused for lane detection. Moving vehicles that have the same colours as the lane markings are eliminated from the feature points before the authors perform angles of inclination and turning points searching to locate the lane boundaries. For unstructured roads, mean-shift segmentation is applied to divide the scene into regions. Possible candidate pairs for road boundaries are elected from the region boundaries, and Bayes rule is used to choose the most probable candidate pairs as the lane boundaries. The experimental results have shown that the classification mechanism can effectively choose the correct lane detection algorithm according to the current environment setting, and the system is able to robustly find the lane boundaries on different types of roads in various weather conditions.