Real time detection of lane markers in urban streets

By Mohamed Aly. IEEE Intelligent Vehicles Symposium, 2008

888 citations recorded by the source corpus.

We present a robust and real time approach to lane marker detection in urban streets. It is based on generating a top view of the road, filtering using selective oriented Gaussian filters, using RANSAC line fitting to give initial guesses to a new and fast RANSAC algorithm for fitting Bezier Splines, which is then followed by a post-processing step. Our algorithm can detect all lanes in still images of the street in various conditions, while operating at a rate of 50 Hz and achieving comparable results to previous techniques.

Subfields: Automotive Engineering, Computer Vision and Pattern Recognition, Environmental Engineering

View Paper: https://www.semanticscholar.org/paper/e6c40fd5adf616b31766f7578fd8c18e18158b78

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Automotive Engineering

In Computer Vision and Pattern Recognition

In Environmental Engineering

Citations in the corpus

Robust Lane Detection From Continuous Driving Scenes Using Deep Neural Networks (2019, methodology): “1) Geometric modeling: Most methods in this group follow a two-step solution, with edge detection at first and line fitting at second [4], [28]–[30].”
VPGNet: Vanishing Point Guided Network for Lane and Road Marking Detection and Recognition (2017, methodology): “For example, Caltech Lanes Dataset [1] contains 1,225 images taken from four different places.”
Spatial As Deep: Spatial CNN for Traffic Scene Understanding (2017, methodology): “based methods, and also gives better results than the much deeper ResNet-101 (He et al. 2016). Related Work. For lane detection, most existing algorithms are based on hand-crafted low-level features (Aly 2008; Son et al. 2015; Jung, Youn, and Sull 2016), limiting there capability to deal with harsh ”
Overview of Environment Perception for Intelligent Vehicles (2017, methodology): “In the RANSAC line fitting-based method [56], a top view of”
Inferring The Latent Structure of Human Decision-Making from Raw Visual Inputs (2017, background): “Mediated perception is a two-step approach that ﬁrst obtains scene information and then makes a driving decision (Aly, 2008; Lenz et al., 2011); behavior reﬂex , on the other hand, adopts a direct approach by mapping visual inputs to driving actions (Pomerleau, 1989; 1991).”
Efficient deep models for monocular road segmentation (2016, background): “Aly et al. [2] use the Hough transformation to identify lane markings and to localize the road area.”
Towards End-to-End Lane Detection: an Instance Segmentation Approach (2018, background): “Popular models are cubic polynomials [32], [25], splines [1] or clothoids [10].”
BDD100K: A Diverse Driving Video Database with Scalable Annotation Tooling (2018, background): “Caltech Lanes Dataset [2] - 1,224 4 1 1 2 Road Marking Dataset [21] - 1,443 29 2 3 10 KITTI-ROAD [11] 289 579 1 1 2 VPGNet [13] 14,783 21,097 4 2 17 Ours 70,000 100,000 100,000 6 3 11 Table 2: Comparisons with other lane marking datasets.”
InfoGAIL: Interpretable Imitation Learning from Visual Demonstrations (2017, background): “Mediated perception is a two-step approach that ﬁrst obtains scene information and then makes a driving decision [34–36]; behavior reﬂex , on the other hand, adopts a direct approach by mapping visual inputs to driving actions [37, 16].”
Rethinking Self-driving: Multi-task Knowledge for Better Generalization and Accident Explanation Ability (2018, background): “For separate scene understanding task, car detection(Lenz et al., 2011) and lane detection(Aly, 2008) are two main topics in this area.”
Deep Neural Network for Structural Prediction and Lane Detection in Traffic Scene (2017, methodology): “We also include the data set in [49] in this test (Caltech data set).”
Approximate Bayesian Image Interpretation using Generative Probabilistic Graphics Programs (2013, neutral): “size and rotation.”
Deep Steering: Learning End-to-End Driving Model from Spatial and Temporal Visual Cues (2017, background): “D orientation), are regarded as key enablers for autonomous driving, including vehicle / pedestrian detection [13], [14], [15], [16] and lane detection [17], [18], [19].”
BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learning (2018, methodology): “For example, the Caltech Lanes Dataset [4] only contains 1,224 images, and the Road Marking Dataset [32] has 1,443 images labeled in 11 classes of lane markings.”
Line-CNN: End-to-End Traffic Line Detection With Line Proposal Unit (2020, methodology): “The first one is a classical method [6] (dubbed “Gaussian+RANSAC”) which deploys Gaussian filters to extract features and then uses a RANSAC [46] spline fitting technique to locate traffic lines.”
DeepDriving: Learning Affordance for Direct Perception in Autonomous Driving (2015, methodology): “Typical algorithms output bounding boxes on detected cars [4, 13] and splines on detected lane markings [1].”
D Traffic Scene Understanding from Movable Platforms (2013, background): “Motivated by this success, several efforts have been conducted towards autonomous driving in real traffic [10], [29], [51] triggering research on lane estimation [3],”
3D Traffic Scene Understanding From Movable Platforms (2014, background): “While autonomous driving on highways has been successfully demonstrated for several decades [21], navigating urban environments remains an unsolved problem.”