SFNet-N: An Improved SFNet Algorithm for Semantic Segmentation of Low-Light Autonomous Driving Road Scenes

Wang, Hai; Chen, Yanyan; Cai, Yingfeng; Chen, Long; Li, Yicheng; Sotelo, Miguel Ángel; Li, Zhixiong

doi:10.1109/tits.2022.3177615

Cited by 102 publications

(35 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Existing LiDAR point cloud moving object segmentation networks can be categorized into two groups: computer-vision-based [ 9 , 10 , 11 , 12 , 13 ] and LiDAR-sensor-based [ 14 , 15 , 16 ]. However, the processing of LiDAR data remains challenging due to the uneven distribution and sparsity of LiDAR point clouds.…”

Section: Related Workmentioning

confidence: 99%

Real-Time LiDAR Point-Cloud Moving Object Segmentation for Autonomous Driving

Xie

Wei

Yang

2023

Sensors

View full text Add to dashboard Cite

The key to autonomous navigation in unmanned systems is the ability to recognize static and moving objects in the environment and to support the task of predicting the future state of the environment, avoiding collisions, and planning. However, because the existing 3D LiDAR point-cloud moving object segmentation (MOS) convolutional neural network (CNN) models are very complex and have large computation burden, it is difficult to perform real-time processing on embedded platforms. In this paper, we propose a lightweight MOS network structure based on LiDAR point-cloud sequence range images with only 2.3 M parameters, which is 66% less than the state-of-the-art network. When running on RTX 3090 GPU, the processing time is 35.82 ms per frame and it achieves an intersection-over-union(IoU) score of 51.3% on the SemanticKITTI dataset. In addition, the proposed CNN successfully runs the FPGA platform using an NVDLA-like hardware architecture, and the system achieves efficient and accurate moving-object segmentation of LiDAR point clouds at a speed of 32 fps, meeting the real-time requirements of autonomous vehicles.

show abstract

Section: Related Workmentioning

confidence: 99%

Real-Time LiDAR Point-Cloud Moving Object Segmentation for Autonomous Driving

Xie

Wei

Yang

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…In agricultural fields [ 3 , 4 , 5 ], semantic segmentation of remote sensing images helps to map and monitor land use and land cover (LULC) changes for sustainable land development, planning, and management. In the autonomous driving system applications [ 6 , 7 ], vehicles need the semantic information of the surrounding scene to assist their understanding and perception of complex traffic situations. In this way, the vehicles can identify lane markings, traffic signs, traffic lights, and other generic objects, thereby achieving real-time lane-level positioning and navigation in normal driving conditions.…”

Section: Introductionmentioning

confidence: 99%

EDPNet: An Encoding–Decoding Network with Pyramidal Representation for Semantic Image Segmentation

Chen

et al. 2023

Sensors

View full text Add to dashboard Cite

This paper proposes an encoding–decoding network with a pyramidal representation module, which will be referred to as EDPNet, and is designed for efficient semantic image segmentation. On the one hand, during the encoding process of the proposed EDPNet, the enhancement of the Xception network, i.e., Xception+ is employed as a backbone to learn the discriminative feature maps. The obtained discriminative features are then fed into the pyramidal representation module, from which the context-augmented features are learned and optimized by leveraging a multi-level feature representation and aggregation process. On the other hand, during the image restoration decoding process, the encoded semantic-rich features are progressively recovered with the assistance of a simplified skip connection mechanism, which performs channel concatenation between high-level encoded features with rich semantic information and low-level features with spatial detail information. The proposed hybrid representation employing the proposed encoding–decoding and pyramidal structures has a global-aware perception and captures fine-grained contours of various geographical objects very well with high computational efficiency. The performance of the proposed EDPNet has been compared against PSPNet, DeepLabv3, and U-Net, employing four benchmark datasets, namely eTRIMS, Cityscapes, PASCAL VOC2012, and CamVid. EDPNet acquired the highest accuracy of 83.6% and 73.8% mIoUs on eTRIMS and PASCAL VOC2012 datasets, while its accuracy on the other two datasets was comparable to that of PSPNet, DeepLabv3, and U-Net models. EDPNet achieved the highest efficiency among the compared models on all datasets.

show abstract

“…A major part of driving behaviour is related to acceleration choice. To this end, many studies have been done on acceleration behaviour and several models have been proposed with different levels of complexity to depict the underlying processes of acceleration decision‐making [8–22, 84, 85].…”

Section: Introductionmentioning

confidence: 99%

A viscous continuum traffic flow model based on the cooperative car‐following behaviour of connected and autonomous vehicles

Jafaripournimchahi

Cai

Wang

et al. 2022

IET Intelligent Trans Sys

Self Cite

View full text Add to dashboard Cite

Connected and Autonomous Vehicles (CAVs) can receive various information from surrounding vehicles through Vehicle‐to‐Everything (V2X) communication technologies and adjust their car‐following behaviour accordingly. Although several studies have evaluated the impact of CAVs on traffic flow stability in a small segment of networks, most approaches are focused on their specific applications considering the trajectory information, and there is a lack of studies analyzing the impact of CAVs on a large‐scale network. This paper proposes a novel viscous continuum traffic model considering the anticipation of space headway, the throttle angle, and brake torque information during cooperative car‐following. The methods employed to develop the new car‐following model and its counterpart continuum traffic model have been described. The linear and non‐linear stability analyses of the newly developed model have been conducted to obtain the critical stability factors in small perturbations. Numerical simulations have been carried out to investigate the effect of the anticipation, the throttle angle, and brake torque information on traffic stability, fuel consumption, and exhaust emissions. The numerical results reveal that the anticipation of space headway and the transmission of the throttle angle and brake torque information during cooperative car‐following manoeuvres can improve the traffic flow stability and reduce fuel consumption and emissions.

show abstract

SFNet-N: An Improved SFNet Algorithm for Semantic Segmentation of Low-Light Autonomous Driving Road Scenes

Cited by 102 publications

References 39 publications

Real-Time LiDAR Point-Cloud Moving Object Segmentation for Autonomous Driving

Real-Time LiDAR Point-Cloud Moving Object Segmentation for Autonomous Driving

EDPNet: An Encoding–Decoding Network with Pyramidal Representation for Semantic Image Segmentation

A viscous continuum traffic flow model based on the cooperative car‐following behaviour of connected and autonomous vehicles

Contact Info

Product

Resources

About