Complex Traffic Scenes Research Articles

Synchronous End-to-End Vehicle Pedestrian Detection Algorithm Based on Improved YOLOv8 in Complex Scenarios.

In modern urban traffic, vehicles and pedestrians are fundamental elements in the study of traffic dynamics. Vehicle and pedestrian detection have significant practical value in fields like autonomous driving, traffic management, and public security. However, traditional detection methods struggle in complex environments due to challenges such as varying scales, target occlusion, and high computational costs, leading to lower detection accuracy and slower performance. To address these challenges, this paper proposes an improved vehicle and pedestrian detection algorithm based on YOLOv8, with the aim of enhancing detection in complex traffic scenes. The motivation behind our design is twofold: first, to address the limitations of traditional methods in handling targets of different scales and severe occlusions, and second, to improve the efficiency and accuracy of real-time detection. The new generation of dense pedestrian detection technology requires higher accuracy, less computing overhead, faster detection speed, and more convenient deployment. Based on the above background, this paper proposes a synchronous end-to-end vehicle pedestrian detection algorithm based on improved YOLOv8, aiming to solve the detection problem in complex scenes. First of all, we have improved YOLOv8 by designing a deformable convolutional improved backbone network and attention mechanism, optimized the network structure, and improved the detection accuracy and speed. Secondly, we introduced an end-to-end target search algorithm to make the algorithm more stable and accurate in vehicle and pedestrian detection. The experimental results show that, using the algorithm designed in this paper, our model achieves an 11.76% increase in precision and a 6.27% boost in mAP. In addition, the model maintains a real-time detection speed of 41.46 FPS, ensuring robust performance even in complex scenarios. These optimizations significantly enhance both the efficiency and robustness of vehicle and pedestrian detection, particularly in crowded urban environments. We further apply our improved YOLOv8 model for real-time detection in intelligent transportation systems and achieve exceptional performance with a mAP of 95.23%, outperforming state-of-the-art models like YOLOv5, YOLOv7, and Faster R-CNN.

Res2Net-based multi-scale and multi-attention model for traffic scene image classification.

With the increasing applications of traffic scene image classification in intelligent transportation systems, there is a growing demand for improved accuracy and robustness in this classification task. However, due to weather conditions, time, lighting variations, and annotation costs, traditional deep learning methods still have limitations in extracting complex traffic scene features and achieving higher recognition accuracy. The previous classification methods for traffic scene images had gaps in multi-scale feature extraction and the combination of frequency domain, spatial, and channel attention. To address these issues, this paper proposes a multi-scale and multi-attention model based on Res2Net. Our proposed framework introduces an Adaptive Feature Refinement Pyramid Module (AFRPM) to enhance multi-scale feature extraction, thus improving the accuracy of traffic scene image classification. Additionally, we integrate frequency domain and spatial-channel attention mechanisms to develop recognition capabilities for complex backgrounds, objects of different scales, and local details in traffic scene images. The paper conducts the task of classifying traffic scene images using the Traffic-Net dataset. The experimental results demonstrate that our model achieves an accuracy of 96.88% on this dataset, which is an improvement of approximately 2% compared to the baseline Res2Net network. Furthermore, we validate the effectiveness of the proposed modules through ablation experiments.

Counting dense object of multiple types based on feature enhancement.

Accurately counting the number of dense objects in an image, such as pedestrians or vehicles, is a challenging and practical task. The existing density map regression methods based on CNN are mainly used to count a class of dense objects in a single scene. However, in complex traffic scenes, objects such as vehicles and pedestrians usually exist at the same time, and multiple classes of dense objects need to be counted simultaneously. To solve the above issues, we propose a new multiple types of dense object counting method based on feature enhancement, which can enhance the features of dense counting objects in complex traffic scenes to realize the classification and regression counting of dense vehicles and people. The counting model consists of the regression subnet and the classification subnet. The regression subnet is primarily used to generate two-channel predicted density maps, mainly including the initial feature layer and the feature enhancement layer, in which the feature enhancement layer can enhance the classification features and regression counting features of dense objects in complex traffic scenes. The classification subnet mainly supervises classifying dense vehicles and people into two feature channels to assist the regression counting task of the regression subnets. Our method is compared on VisDrone+ datasets, ApolloScape+ datasets, and UAVDT+ datasets. The experimental results show that the method counts two kinds of dense objects simultaneously and outputs a high-quality two-channel predicted density map. The counting performance is better than the state-of-the-art counting network in dense people and vehicle counting. In future work, we will further improve the feature extraction ability of the model in complex traffic scenes to classify and count a variety of dense objects such as cars, pedestrians, and non-motor vehicles.

PDT-YOLO: A Roadside Object-Detection Algorithm for Multiscale and Occluded Targets.

To tackle the challenges of weak sensing capacity for multi-scale objects, high missed detection rates for occluded targets, and difficulties for model deployment in detection tasks of intelligent roadside perception systems, the PDT-YOLO algorithm based on YOLOv7-tiny is proposed. Firstly, we introduce the intra-scale feature interaction module (AIFI) and reconstruct the feature pyramid structure to enhance the detection accuracy of multi-scale targets. Secondly, a lightweight convolution module (GSConv) is introduced to construct a multi-scale efficient layer aggregation network module (ETG), enhancing the network feature extraction ability while maintaining weight. Thirdly, multi-attention mechanisms are integrated to optimize the feature expression ability of occluded targets in complex scenarios, Finally, Wise-IoU with a dynamic non-monotonic focusing mechanism improves the accuracy and generalization ability of model sensing. Compared with YOLOv7-tiny, PDT-YOLO on the DAIR-V2X-C dataset improves mAP50 and mAP50:95 by 4.6% and 12.8%, with a parameter count of 6.1 million; on the IVODC dataset by 15.7% and 11.1%. We deployed the PDT-YOLO in an actual traffic environment based on a robot operating system (ROS), with a detection frame rate of 90 FPS, which can meet the needs of roadside object detection and edge deployment in complex traffic scenes.

YCANet: Target Detection for Complex Traffic Scenes Based on Camera-LiDAR Fusion

YCANet: Target Detection for Complex Traffic Scenes Based on Camera-LiDAR Fusion

Processing of complex traffic scenes for effective steering and collision avoidance: a perspective, from research into human control, on the challenges for sensor-based autonomous vehicles on urban roads.

An overview is provided of behavioral research into human steering and collision avoidance including the processing of optic flow, optical looming and the role of the human mobile gaze system. A consideration is then made of the issues that may occur for autonomous vehicles (AV) when they move from grid-type road networks into complex inner-city streets and interact with human drivers, pedestrians and cyclists. Comparisons between human processing and AV processing of these interactions are made. This raises issues as to whether AV control systems need to mimic human visual processing more closely and highlights the need for AV systems to develop a "theory of road users" that allows attribution of intent to other drivers, cyclists or pedestrians. Guidelines for the development of a "theory of road users" for AVs are suggested.

An Interactively Motion-Assisted Network for Multiple Object Tracking in Complex Traffic Scenes

An Interactively Motion-Assisted Network for Multiple Object Tracking in Complex Traffic Scenes

Vehicle recognition pipeline via DeepSort on aerial image datasets.

Unmanned aerial vehicles (UAVs) are widely used in various computer vision applications, especially in intelligent traffic monitoring, as they are agile and simplify operations while boosting efficiency. However, automating these procedures is still a significant challenge due to the difficulty of extracting foreground (vehicle) information from complex traffic scenes. This paper presents a unique method for autonomous vehicle surveillance that uses FCM to segment aerial images. YOLOv8, which is known for its ability to detect tiny objects, is then used to detect vehicles. Additionally, a system that utilizes ORB features is employed to support vehicle recognition, assignment, and recovery across picture frames. Vehicle tracking is accomplished using DeepSORT, which elegantly combines Kalman filtering with deep learning to achieve precise results. Our proposed model demonstrates remarkable performance in vehicle identification and tracking with precision of 0.86 and 0.84 on the VEDAI and SRTID datasets, respectively, for vehicle detection. For vehicle tracking, the model achieves accuracies of 0.89 and 0.85 on the VEDAI and SRTID datasets, respectively.

End-to-End Object Detection by Sparse R-CNN With Hybrid Matching in Complex Traffic Scenes

End-to-End Object Detection by Sparse R-CNN With Hybrid Matching in Complex Traffic Scenes

A Feature Fusion Network with Multiscale Adaptively Attentional for Object Detection in Complex Traffic Scenes

A Feature Fusion Network with Multiscale Adaptively Attentional for Object Detection in Complex Traffic Scenes

SASAN: Shape-Adaptive Set Abstraction Network for Point-Voxel 3D Object Detection.

Point-voxel 3D object detectors have achieved impressive performance in complex traffic scenes. However, they utilize the 3D sparse convolution (spconv) layers with fixed receptive fields, such as voxel-based detectors, and inherit the fixed sphere radius from point-based methods for generating the features of keypoints, which make them weak in adaptively modeling various geometrical deformations and sizes of real objects. To tackle this issue, we propose a shape-adaptive set abstraction network (SASAN) for point-voxel 3D object detection. First, the proposal and offset generation module is adopted to learn the coordinates and confidences of 3D proposals and shape-adaptive offsets of the certain number of offset points for each voxel. Meanwhile, an extra offset supervision task is employed to guide the learning of shifting values of offset points, aiming at motivating the predicted offsets to preferably adapt to the various shapes of objects. Then, the shape-adaptive set abstraction module is proposed to extract multiscale keypoints features by grouping the neighboring offset points' features, as well as features learned from adjacent raw points and the 2-D bird-view map. Finally, the region of interest (RoI)-grid proposal refinement module is used to aggregate the keypoints features for further proposal refinement and confidence prediction. Extensive experiments on the competitive KITTI 3D detection benchmark demonstrate that the proposed SASAN gains superior performance as compared with state-of-the-art methods.

HCLT-YOLO: A Hybrid CNN and Lightweight Transformer Architecture for Object Detection in Complex Traffic Scenes

HCLT-YOLO: A Hybrid CNN and Lightweight Transformer Architecture for Object Detection in Complex Traffic Scenes

IMTRAC: Image-Based Modeling for Transformative Traffic Control via Foundation Models

This paper introduces the Image-based Modeling for Transformative Traffic Control (IMTRAC) framework, which revolutionizes urban traffic management by integrating image-based representations, foundation models, and dynamic signaling. It transitions from textual to visual data analysis, transforming complex traffic scenes into quantifiable insights. The image-based representation is then analyzed by foundation models to infer refined and effective traffic control policies. Our experimental results show marked improvements in traffic flow and reductions in waiting times, highlighting IMTRAC’s ability to outperform traditional control methods. The proposed IMTRAC promises a significant leap towards smarter, efficient urban traffic systems, leveraging the synergy of advanced imaging techniques and multi-modal foundation models to enhance control decisions.

A strategy of vehicle following on slope road at night considering the safety of the intended functionality

As hardware devices that help intelligent vehicles perceive the traffic environment, sensors play a crucial role in vehicle safety control. However, in the face of a complex traffic scene, detection failure is a safety of the intended functionality (SOTIF) problems that may occur due to sensors’ limited functions. For example, under slope road at night, the radar sensor may lose the target because of the narrow detection perspective, vision sensors is prone to severe distortion because of poor light. Therefore, this paper proposes a following strategy to address the SOTIF problems like this. By analyzing the historical data, the motion behavior of the preceding vehicle when SOTIF events happen is predicted based on the improved Markov model. Then, the observation and expectation models of the vehicular distance are designed based on vehicle movement and slope conditions to obtain the distance information. The sliding mode controller is designed to control velocity and prevent collisions and other accidents reasonably. Finally, simulation and hardware-in-the-loop (HiL) tests verify that the proposed method can improve vehicle safety when the sensors’ function fails on slope road at night.

A method of knowledge distillation based on feature fusion and attention mechanism for complex traffic scenes

Object detectors based on deep learning can run smoothly on a terminal device in complex traffic scenes, and the model compression method has become a research hotspot. Considering student network single learning in the knowledge distillation algorithm, the dependence on loss function design leads to parameter sensitivity and other problems, we propose a new knowledge distillation method with second-order term attention mechanisms and feature fusion of adjacent layers. First, we build a knowledge distillation framework based on YOLOv5 and propose a new attention mechanism in the teacher network backbone to extract the hot map. Then, we combine the hot map features with the next level features through the fusion module. By fusing the useful information of the low convolution layer and the feature map of the high convolution layer to help the student network obtain the final prediction map. Finally, to improve the accuracy of small objects, we add a 160 × 160 detection head and use a transformer encoder block module to replace the convolution network of the head. Sufficient experimental results show that our method achieves state-of-the-art performance. The speed and number of parameters remain unchanged, but the average detection accuracy is 97.4% on the KITTI test set. On the Cityscapes test set, the average detection accuracy reaches 92.7%.

Semi-supervised lane detection for continuous traffic scenes

Objective This article aims to upgrade the lane detection algorithm from image to video level in order to advance automatic driving technology. The objective is to propose a cost-efficient algorithm that can handle complex traffic scenes and different driving speeds using continuous image inputs. Methods To achieve this objective, we introduce the Multi-ERFNet-ConvLSTM network framework, which combines Efficient Residual Factorized ConvNet (ERFNet) and Convolution Long Short Term Memory (ConvLSTM). Additionally, we incorporate the Pyramidally Attended Feature Extraction (PAFE) Module into our network design to effectively handle multi-scale lane objects. The algorithm is evaluated using a divided dataset and comprehensive assessments are conducted across multiple dimensions. Results In the testing phase, the Multi-ERFNet-ConvLSTM algorithm surpasses the primary baselines and demonstrates superior performance in terms of Accuracy, Precision, and F1-score metrics. It exhibits excellent detection results in various complex traffic scenes and performs well at different driving speeds. Conclusions The proposed Multi-ERFNet-ConvLSTM algorithm provides a robust solution for video-level lane detection in advanced automatic driving. By utilizing continuous image inputs and incorporating the PAFE Module, the algorithm achieves high performance while reducing labeling costs. Its exceptional accuracy, precision, and F1-score metrics highlight its effectiveness in complex traffic scenarios. Moreover, its adaptability to different driving speeds makes it suitable for real-world applications in autonomous driving systems.

Improved Mask R-CNN Multi-Target Detection and Segmentation for Autonomous Driving in Complex Scenes.

Vision-based target detection and segmentation has been an important research content for environment perception in autonomous driving, but the mainstream target detection and segmentation algorithms have the problems of low detection accuracy and poor mask segmentation quality for multi-target detection and segmentation in complex traffic scenes. To address this problem, this paper improved the Mask R-CNN by replacing the backbone network ResNet with the ResNeXt network with group convolution to further improve the feature extraction capability of the model. Furthermore, a bottom-up path enhancement strategy was added to the Feature Pyramid Network (FPN) to achieve feature fusion, while an efficient channel attention module (ECA) was added to the backbone feature extraction network to optimize the high-level low resolution semantic information graph. Finally, the bounding box regression loss function smooth L1 loss was replaced by CIoU loss to speed up the model convergence and minimize the error. The experimental results showed that the improved Mask R-CNN algorithm achieved 62.62% mAP for target detection and 57.58% mAP for segmentation accuracy on the publicly available CityScapes autonomous driving dataset, which were 4.73% and 3.96%% better than the original Mask R-CNN algorithm, respectively. The migration experiments showed that it has good detection and segmentation effects in each traffic scenario of the publicly available BDD autonomous driving dataset.

Complex Traffic Scene Image Classification Based on Sparse Optimization Boundary Semantics Deep Learning

With the rapid development of intelligent traffic information monitoring technology, accurate identification of vehicles, pedestrians and other objects on the road has become particularly important. Therefore, in order to improve the recognition and classification accuracy of image objects in complex traffic scenes, this paper proposes a segmentation method of semantic redefine segmentation using image boundary region. First, we use the SegNet semantic segmentation model to obtain the rough classification features of the vehicle road object, then use the simple linear iterative clustering (SLIC) algorithm to obtain the over segmented area of the image, which can determine the classification of each pixel in each super pixel area, and then optimize the target segmentation of the boundary and small areas in the vehicle road image. Finally, the edge recovery ability of condition random field (CRF) is used to refine the image boundary. The experimental results show that compared with FCN-8s and SegNet, the pixel accuracy of the proposed algorithm in this paper improves by 2.33% and 0.57%, respectively. And compared with Unet, the algorithm in this paper performs better when dealing with multi-target segmentation.

VV-YOLO: A Vehicle View Object Detection Model Based on Improved YOLOv4

Vehicle view object detection technology is the key to the environment perception modules of autonomous vehicles, which is crucial for driving safety. In view of the characteristics of complex scenes, such as dim light, occlusion, and long distance, an improved YOLOv4-based vehicle view object detection model, VV-YOLO, is proposed in this paper. The VV-YOLO model adopts the implementation mode based on anchor frames. In the anchor frame clustering, the improved K-means++ algorithm is used to reduce the possibility of instability in anchor frame clustering results caused by the random selection of a cluster center, so that the model can obtain a reasonable original anchor frame. Firstly, the CA-PAN network was designed by adding a coordinate attention mechanism, which was used in the neck network of the VV-YOLO model; the multidimensional modeling of image feature channel relationships was realized; and the extraction effect of complex image features was improved. Secondly, in order to ensure the sufficiency of model training, the loss function of the VV-YOLO model was reconstructed based on the focus function, which alleviated the problem of training imbalance caused by the unbalanced distribution of training data. Finally, the KITTI dataset was selected as the test set to conduct the index quantification experiment. The results showed that the precision and average precision of the VV-YOLO model were 90.68% and 80.01%, respectively, which were 6.88% and 3.44% higher than those of the YOLOv4 model, and the model's calculation time on the same hardware platform did not increase significantly. In addition to testing on the KITTI dataset, we also selected the BDD100K dataset and typical complex traffic scene data collected in the field to conduct a visual comparison test of the results, and then the validity and robustness of the VV-YOLO model were verified.

A Novel Method for Obstacle Detection in Front of Vehicles Based on the Local Spatial Features of Point Cloud

Obstacle detection is the primary task of the Advanced Driving Assistance System (ADAS). However, it is very difficult to achieve accurate obstacle detection in complex traffic scenes. To this end, this paper proposes an obstacle detection method based on the local spatial features of point clouds. Firstly, the local spatial point cloud of a superpixel is obtained through stereo matching and the SLIC image segmentation algorithm. Then, the probability of the obstacle in the corresponding area is estimated from the spatial feature information of the local plane normal vector and the superpixel point-cloud height, respectively. Finally, the detection results of the two methods are input into the Bayesian framework in the form of probabilities for the final decision. In order to describe the traffic scene efficiently and accurately, the detection results are further transformed into a multi-layer stixel representation. We carried out experiments on the KITTI dataset and compared several obstacle detection methods. The experimental results indicate that the proposed method has advantages in terms of its Pixel-wise True Positive Rate (PTPR) and Pixel-wise False Positive Rate (PFPR), particularly in complex traffic scenes, such as uneven roads.