Awesome-Lane-Detection

本仓库由公众号【自动驾驶之心】 团队整理，欢迎关注，一览最前沿的技术分享！

自动驾驶之心是国内首个自动驾驶开发者社区！这里有最全面有效的自动驾驶与AI学习路线（感知/定位/融合）和自动驾驶与AI公司内推机会！

一、Based on detection solutions

1.Anchor-based Classic Articles

CLRNet：Cross Layer Refinement Network for Lane Detection

CondLaneNet：a Top-to-down Lane Detection Framework Based on Conditional Convolution

End-to-End Lane Marker Detection via Row-wise Classification

FastDraw：Addressing the Long Tail of Lane Detection by Adapting a Sequential Prediction Network、

Keep your Eyes on the Lane: Real-time Attention-guided Lane Detection

Structure Guided Lane Detection

UFLD V1：Ultra Fast Structure-aware Deep Lane Detection

UFLD v2：Ultra Fast Deep Lane Detection with Hybrid Anchor Driven Ordinal Classification

2. Detection-based approach

CLRNet：Cross Layer Refinement Network for Lane Detection

Keep your Eyes on the Lane： Real-time Attention-guided Lane Detection

Keep your Eyes on the Lane：Attention-guided Lane Detection

3.DILane: Dynamic Instance-Aware Network for Lane Detection

4.RoadSegNet: a deep learning framework for autonomous urban road detection

二、Segmentation based

1. Semantic segmentation-based lane line detection

EL-GAN：Embedding Loss Driven Generative Adversarial Networks for Lane Detection

Learning Lightweight Lane Detection CNNs by Self Attention Distillation

Inter-Region Affinity Distillation for Road Marking Segmentation

Towards End-to-End Lane Detection: an Instance Segmentation Approach

Polylanenet：Lane estimation via deep polynomial regression

RESA：Recurrent Feature-Shift Aggregator for Lane Detection

SCNN：Spatial As Deep_ Spatial CNN for Traffic Scene Understanding

2. Segmentation-based approach

ContinuityLearner：Geometric Continuity Feature Learning for Lane Segmentation

Eigenlanes：Data-Driven Lane Descriptors for Structurally Diverse Lanes

Lane detection with Position Embedding

Lane Detection with Versatile AtrousFormer and Local Semantic Guidance

LaneAF：Robust Multi-Lane Detection with Affinity Fields

Multi-Class Lane Semantic Segmentation using Efficient Convolutional Networks

3. Segmentation-based approach

Multi-lane Detection Using Instance Segmentation and Attentive Voting

Multi-level Domain Adaptation for Lane Detection

Structure Guided Lane Detection

Towards End-to-End Lane Detection：an Instance Segmentation approach

Towards Lightweight Lane Detection by Optimizing Spatial Embedding

Ultra Fast Structure-aware Deep Lane Detection

4.RoadSegNet: a deep learning framework for autonomous urban road detection

三、Based on the classification scheme

End-to-End Lane Marker Detection via Row-wise Classification

SwiftLane：Towards Fast and Efficient Lane Detection

Ultra Fast Deep Lane Detection with Hybrid Anchor Driven Ordinal Classification

四、Curve-based prediction schemes

1. Based on curve fitting method

End-to-end Lane Detection through Differentiable Least-Squares Fitting

Rethinking Efficient Lane Detection via Curve Modeling

五、Key point based forecasting solution

1. Classic Articles

FOLOLane：Focus on Local：Detecting Lane Marker from Bottom Up via Key Point

GANet：A Keypoint-based Global Association Network for Lane Detection

End-to-end Lane Shape Prediction with Transformers

PINet：Key Points Estimation and Point Instance Segmentation Approach for Lane Detection

六、Stereo based

1. Stereo based

Real-Time Stereo Vision-Based Lane Detection system

七、Sequence-based

A Hybrid Spatial-temporal Sequence-to-one Neural Network Model for Lane Detection

LaneNet：Real-Time Lane Detection Networks for Autonomous Driving

RESA：Recurrent Feature-Shift Aggregator for Lane Detection

Robust Lane Detection from Continuous Driving Scenes Using Deep Neural Networks

VIL-100：A New Dataset and A Baseline Model for Video Instance Lane Detection

八、Transformer-based

Laneformer：Object-aware Row-Column Transformers for Lane Detection

Multitasking Learning

1.多任务学习经典文章YOLOP | YOLOP

YOLOP：You Only Look Once for Panoptic Driving Perception

十、Multi-sensor based

Deep Multi-Sensor Lane Detection

FusionLane：Multi-Sensor Fusion for Lane Marking

十一 、LiDAR based

1.Extraction of lane markings based on 64-lane LiDAR reflectance data

Road Markings Segmentation from LIDAR Point Clouds using Reflectivity Information

十二、Other

Deep Multi-Sensor Lane Detection

FusionLane：Multi-Sensor Fusion for Lane Marking

十三、3D lane lines

3D-LaneNet：End-to-End 3D Multiple Lane Detection

3D-LaneNet+：Anchor Free Lane Detection using a Semi-Local Representation

ONCE-3DLanes：Building Monocular 3D Lane Detection

PersFormer：3D Lane Detection via Perspective Transformer and the OpenLane Benchmark

Semi-Local 3D Lane Detection and Uncertainty estimation

2.M2-3DLaneNet

M2-3DLaneNet: Multi-Modal 3D Lane Detection

[Code]

3.BEV-LaneDet

BEV-LaneDet: Fast Lane Detection on BEV Ground

[Code]

4.OpenLanes Dataset

[Dataset]

十四、Summary of lane line identification issues

遮挡、磨损、不连续：解决方案包括拟合估计、结合地图等

Occlusion, wear, discontinuity: solutions include fitting estimates, combining maps, etc.

细长结构、需要捕捉细节特征：通过分割模型优化等方法

Slender structure, need to capture detailed features: by segmentation model optimization and other methods

十五、Code

1. Lane Line Detection

[Ultra_Fast_Lane_Detection_TensorRT]

[Ultra-Fast-Lane-Detection-V2]

[LaneNet-Lane-Detection]

[Codes-for-Lane-Detection]

[Lane Finding Project for Self-Driving Car ND]

[Spatial CNN for Traffic Lane Detection]

[End-to-end Lane Detection]

[LaneATT]

[Advanced Lane Detection]

[VPGNet Usage and Lane Detection]

[AMREL]

2. Multitasking Methods

[YOLOP-TensorRT]

[YOLOPv2]