Search Results (716)

To address the issues of insufficient samples, limited scene diversity, missing perspectives, and low resolution in existing UAV-based pedestrian detection datasets, this paper proposes a novel UAV-based pedestrian detection benchmark dataset named the Novel Surveillance View (NSV). This dataset encompasses diverse scenes and pedestrian information captured from multiple perspectives, and introduces an innovative data mining approach that leverages tracking and optical flow information. This approach significantly improves data acquisition efficiency while ensuring annotation quality. Furthermore, an improved pedestrian detection method is proposed to overcome the performance degradation caused by significant perspective changes in top-down UAV views. Firstly, the View-Agnostic Decomposition (VAD) module decouples features into perspective-dependent and perspective-independent branches to enhance the model’s generalization ability to perspective variations. Secondly, the Deformable Conv-BN-SiLU (DCBS) module dynamically adjusts the receptive field shape to better adapt to the geometric deformations of pedestrians. Finally, the Context-Aware Pyramid Spatial Attention (CPSA) module integrates multi-scale features with attention mechanisms to address the challenge of drastic target scale variations. The experimental results demonstrate that the proposed method improves the mean Average Precision (mAP) by 9% on the NSV dataset, thereby validating that the approach effectively enhances pedestrian detection accuracy from UAV perspectives by optimizing perspective features. Full article

This article examines the components of active pedestrian crossing systems, focusing on an innovative solution implemented at Jaworski Street in Kielce. The proposed method for evaluating the effectiveness of this system is based on observations and performance analysis under conditions of limited visibility and increased pedestrian traffic. The key elements of the methodology include assessing the accuracy of pedestrian detection using IR/MW and curtain sensors, analyzing system errors such as false activations or failures to detect pedestrians, and evaluating driver responses to warning signals, with an emphasis on optimizing signal activation timing. The innovation of the method lies in its flexibility and real-time adaptability to dynamic traffic conditions, significantly enhancing system performance and pedestrian safety at crossings. Through integrated detection and activation time management mechanisms, the system can respond to changing conditions in real time, ensuring optimal operational effectiveness. Full article

Object detection in images is a fundamental component of many safety-critical systems, such as autonomous driving, video surveillance systems, and robotics. Adversarial patch attacks, being easily implemented in the real world, provide effective counteraction to object detection by state-of-the-art neural-based detectors. It poses a serious danger in various fields of activity. Existing defense methods against patch attacks are insufficiently effective, which underlines the need to develop new reliable solutions. In this manuscript, we propose a method which helps to increase the robustness of neural network systems to the input adversarial images. The proposed method consists of a Deep Convolutional Neural Network to reconstruct a benign image from the adversarial one; a Calculating Maximum Error block to highlight the mismatches between input and reconstructed images; a Localizing Anomalous Fragments block to extract the anomalous regions using the Isolation Forest algorithm from histograms of images’ fragments; and a Clustering and Processing block to group and evaluate the extracted anomalous regions. The proposed method, based on anomaly localization, demonstrates high resistance to adversarial patch attacks while maintaining the high quality of object detection. The experimental results show that the proposed method is effective in defending against adversarial patch attacks. Using the YOLOv3 algorithm with the proposed defensive method for pedestrian detection in the INRIAPerson dataset under the adversarial attacks, the mAP50 metric reaches 80.97% compared to 46.79% without a defensive method. The results of the research demonstrate that the proposed method is promising for improvement of object detection systems security. Full article

This paper explores the development and testing of two Internet of Things (IoT) applications designed to leverage Vehicle-to-Infrastructure (V2I) communication for managing intelligent intersections. The first scenario focuses on enabling the rapid and safe passage of emergency vehicles through intersections by notifying approaching drivers via a mobile application. The second scenario enhances pedestrian safety by alerting drivers, through the same application, about the presence of pedestrians detected at crosswalks by a traffic sensor equipped with neural network capabilities. Both scenarios were tested at two distinct intelligent intersections in Lioni, Avellino, Italy, and demonstrated notable effectiveness. Results show a significant reduction in emergency vehicle response times and a measurable increase in driver awareness of pedestrians at crossings. The findings underscore the potential of V2I technologies to improve traffic flow, reduce risks for vulnerable road users, and contribute to the advancement of safer and smarter urban transportation systems. Full article

Pedestrian detection is widely used in real-time surveillance, urban traffic, and other fields. As a crucial direction in pedestrian detection, dense pedestrian detection still faces many unresolved challenges. Existing methods suffer from low detection accuracy, high miss rates, large model parameters, and poor robustness. In this paper, to address these issues, we propose a lightweight dense pedestrian detection model with finer-grained feature information interaction called MSCD-YOLO, which can achieve high accuracy, high performance and robustness with only a small number of parameters. In our model, the light-weight backbone network MobileViT is used to reduce the number of parameters while efficiently extracting both local and global features; the SCNeck neck network is designed to fuse the extracted features without losing information; and the DEHead detection head is utilized for multi-scale feature fusion to detect the targets. To demonstrate the effectiveness of our model, we conducted tests on the highly challenging dense pedestrian detection datasets Crowdhuman and Widerperson. Compared to the baseline model YOLOv8n, MSCD-YOLO achieved a 4.6% and 1.8% improvement in [email protected], and a 5.3% and 2.6% improvement in [email protected]:0.95 on the Crowdhuman and Widerperson datasets, respectively. The experimental results show that under the same experimental conditions, MSCD-YOLO significantly outperforms the original model in terms of detection accuracy, efficiency, and model complexity. Full article

Pedestrian safety is increasingly becoming a major concern within and around intersections. This paper outlines a novel approach to enhancing pedestrian safety using a wearable augmented reality (AR) system integrated with live roadside light detection and range (LiDAR) sensor data. The proposed system aims to provide real-time and spatially located warnings to pedestrians, thereby helping them proactively evade potential accidents. The system architecture is built to lessen the burden on edge devices, such as the AR headset itself, and ensure that most of the processor-heavy computations are performed within the server. The effectiveness of this system has been demonstrated and evaluated through various tests, including latency measurements. Point cloud accuracy measurements were recorded with an average offset of approximately 0.73 inches between a conflicting vehicle’s actual location and the visualized object location through AR. This paper also discusses the potential of this system with respect to vehicle-to-vehicle (V2V) systems and other societal benefits. Full article

Autonomous driving (AD) technology has seen significant advancements in recent years; however, challenges remain, particularly in achieving reliable performance under adverse weather conditions such as heavy fog. In response, we propose a multi-class fog density classification approach to enhance the AD system performance. By categorizing fog density into multiple levels (25%, 50%, 75%, and 100%) and generating separate datasets for each class using the CARLA simulator, we improve the perception accuracy for each specific fog density level and analyze the effects of varying fog intensities. This targeted approach offers benefits such as improved object detection, specialized training for each fog class, and increased generalizability. Our results demonstrate enhanced perception of various objects, including cars, buses, trucks, vans, pedestrians, and traffic lights, across all fog densities. This multi-class fog density method is a promising advancement toward achieving reliable AD performance in challenging weather, improving both the precision and recall of object detection algorithms under diverse fog conditions. Full article

As the demand for autonomous driving (AD) systems has increased, the enhancement of their safety has become critically important. A fundamental capability of AD systems is object detection and trajectory forecasting of vehicles and pedestrians around the ego-vehicle, which is essential for preventing potential collisions. This study introduces the Deep learning-based Acceleration-aware Trajectory forecasting (DAT) model, a deep learning-based approach for object detection and trajectory forecasting, utilizing raw sensor measurements. DAT is an end-to-end model that processes sequential sensor data to detect objects and forecasts their future trajectories at each time step. The core innovation of DAT lies in its novel forecasting module, which leverages acceleration data to enhance trajectory forecasting, leading to the consideration of a variety of agent motion models. We propose a robust and innovative method for estimating ground-truth acceleration for objects, along with an object detector that predicts acceleration attributes for each detected object and a novel method for trajectory forecasting. DAT is trained and evaluated on the NuScenes dataset, demonstrating its empirical effectiveness through extensive experiments. The results indicate that DAT significantly surpasses state-of-the-art methods, particularly in enhancing forecasting accuracy for objects exhibiting both linear and nonlinear motion patterns, achieving up to a $2\times$ improvement. This advancement highlights the critical role of incorporating acceleration data into predictive models, representing a substantial step forward in the development of safer autonomous driving systems. Full article

Three-dimensional object detection has been a key area of research in recent years because of its rich spatial information and superior performance in addressing occlusion issues. However, the performance of 3D object detection still lags significantly behind that of 2D object detection, owing to challenges such as difficulties in feature extraction and a lack of texture information. To address this issue, this study proposes a 3D object detection network, CaLiJD (Camera and Lidar Joint Contender for 3D object Detection), guided by two-dimensional detection results. CaLiJD creatively integrates advanced channel attention mechanisms with a novel bounding-box filtering method to improve detection accuracy, especially for small and occluded objects. Bounding boxes are detected by the 2D and 3D networks for the same object in the same scene as an associated pair. The detection results that satisfy the criteria are then fed into the fusion layer for training. In this study, a novel fusion network is proposed. It consists of numerous convolutions arranged in both sequential and parallel forms and includes a Grouped Channel Attention Module for extracting interactions among multi-channel information. Moreover, a novel bounding-box filtering mechanism was introduced, incorporating the normalized distance from the object to the radar as a filtering criterion within the process. Experiments were conducted using the KITTI 3D object detection benchmark. The results showed that a substantial improvement in mean Average Precision (mAP) was achieved by CaLiJD compared with the baseline single-modal 3D detection model, with an enhancement of 7.54%. Moreover, the improvement achieved by our method surpasses that of other classical fusion networks by an additional 0.82%. In particular, CaLiJD achieved mAP values of 73.04% and 59.86%, respectively, thus demonstrating state-of-the-art performance for challenging small-object detection tasks such as those involving cyclists and pedestrians. Full article

Accurate multi-object tracking (MOT) is essential for autonomous vehicles, enabling them to perceive and interact with dynamic environments effectively. Single-modality 3D MOT algorithms often face limitations due to sensor constraints, resulting in unreliable tracking. Recent multi-modal approaches have improved performance but rely heavily on complex, deep-learning-based fusion techniques. In this work, we present CLF-BEVSORT, a camera-LiDAR fusion model operating in the bird’s eye view (BEV) space using the SORT tracking framework. The proposed method introduces a novel association strategy that incorporates structural similarity into the cost function, enabling effective data fusion between 2D camera detections and 3D LiDAR detections for robust track recovery during short occlusions by leveraging LiDAR depth. Evaluated on the KITTI dataset, CLF-BEVSORT achieves state-of-the-art performance with a HOTA score of 77.26% for the Car class, surpassing StrongFusionMOT and DeepFusionMOT by 2.13%, with high precision (85.13%) and recall (80.45%). For the Pedestrian class, it achieves a HOTA score of 46.03%, outperforming Be-Track and StrongFusionMOT by (6.16%). Additionally, CLF-BEVSORT reduces identity switches (IDSW) by over 45% for cars compared to baselines AB3DMOT and BEVSORT, demonstrating robust, consistent tracking and setting a new benchmark for 3DMOT in autonomous driving. Full article

In traffic scenes, pedestrian target detection faces significant issues of misdetection and omission due to factors such as crowd density and obstacle occlusion. To address these challenges and enhance detection accuracy, we propose an improved CCW-YOLO algorithm. The algorithm first introduces a lightweight convolutional layer using GhostConv and incorporates an enhanced C2f module to improve the network’s detection performance. Additionally, it integrates the Coordinate Attention module to better capture key points of the targets. Next, the bounding box loss function CIoU loss at the output of YOLOv5 is replaced with WiseIoU loss to enhance adaptability to various detection scenarios, thereby further improving accuracy. Finally, we develop a pedestrian count detection system using PyQt5 to enhance human–computer interaction. Experimental results on the INRIA public dataset showed that our algorithm achieved a detection accuracy of 98.4%, representing a 10.1% improvement over the original YOLOv5s algorithm. This advancement significantly enhances the detection of small objects in images and effectively addresses misdetection and omission issues in complex environments. These findings have important practical implications for ensuring traffic safety and optimizing traffic flow. Full article

According to statistics from the World Health Organization, traffic accidents are one of the leading causes of death among children and young people, and statistical indicators are even worse for the elderly population. Preventive measures require an approach that includes analyses of traffic infrastructure and regulations, users’ traffic behavior, and their interactions. In this study, a methodology based on traffic microsimulations was developed to select the optimal reconstruction solution for urban traffic infrastructure from the perspective of traffic safety. Comprehensive analyses of local traffic conditions at the selected location, infrastructural properties, and properties related to traffic users were carried out. The developed methodology was applied and tested at a selected unsignalized pedestrian crosswalk located in Osijek, Croatia, where traffic safety issues had been detected. Analyses of the possible solutions for traffic safety improvements were carried out, taking into account the specificities of the chosen location and the traffic participants’ behaviors, which were recorded and measured. The statistical analysis showed that children had shorter reaction times and crossed the street faster than the analyzed group of adult pedestrians, which was dominated by elderly people in this case. Using microsimulation traffic modeling (VISSIM), an analysis was conducted on the incoming vehicle speeds for both the existing and the reconstructed conflict zone solutions under different traffic conditions. The results exhibited a decrease in average speeds for the proposed solution, and traffic volume was detected to have a great impact on incoming speeds. The developed methodology proved to be effective in selecting a traffic solution that respects the needs of both motorized traffic and pedestrians. Full article

To improve the accuracy of detecting small and long-distance objects while self-driving cars are in motion, in this paper, we propose a 3D object detection method, Att-BEVFusion, which fuses camera and LiDAR data in a bird’s-eye view (BEV). First, the transformation from the camera view to the BEV space is achieved through an implicit supervision-based method, and then the LiDAR BEV feature point cloud is voxelized and converted into BEV features. Then, a channel attention mechanism is introduced to design a BEV feature fusion network to realize the fusion of camera BEV feature space and LiDAR BEV feature space. Finally, regarding the issue of insufficient global reasoning in the BEV fusion features generated by the channel attention mechanism, as well as the challenge of inadequate interaction between features. We further develop a BEV self-attention mechanism to apply global operations on the features. This paper evaluates the effectiveness of the Att-BEVFusion fusion algorithm on the nuScenes dataset, and the results demonstrate that the algorithm achieved 72.0% mean average precision (mAP) and 74.3% nuScenes detection score (NDS), with an advanced detection accuracy of 88.9% and 91.8% for single-item detection of automotive and pedestrian categories, respectively. Full article

For drone-based detection tasks, accurately identifying small-scale targets like people, bicycles, and pedestrians remains a key challenge. In this paper, we propose DV-DETR, an improved detection model based on the Real-Time Detection Transformer (RT-DETR), specifically optimized for small target detection in high-density scenes. To achieve this, we introduce three main enhancements: (1) ResNet18 as the backbone network to improve feature extraction and reduce model complexity; (2) the integration of recalibration attention units and deformable attention mechanisms in the neck network to enhance multi-scale feature fusion and improve localization accuracy; and (3) the use of the Focaler-IoU loss function to better handle the imbalanced distribution of target scales and focus on challenging samples. Experimental results on the VisDrone2019 dataset show that DV-DETR achieves an [email protected] of 50.1%, a 1.7% improvement over the baseline model, while increasing detection speed from 75 FPS to 90 FPS, meeting real-time processing requirements. These improvements not only enhance the model’s accuracy and efficiency but also provide practical significance in complex, high-density urban environments, supporting real-world applications in UAV-based surveillance and monitoring tasks. Full article

Accidents involving cyclists and trucks are among the most severe road accidents. In 2021, 199 cyclists were killed in accidents involving a truck in the EU. The main accident situation is a truck turning right and a cyclist going straight ahead. A large proportion of these accidents are caused by the inadequate visibility in an HGV (Heavy Goods Vehicle). The blind spot, in particular, is a significant contributor to these accidents. A BSD (Blind Spot Detection) system is expected to significantly reduce these accidents. There are only a few studies that estimate the potential of assistance systems, and these studies include a combined assessment of cyclists and pedestrians. In the present study, accident simulations are used to assess a warning and an autonomously intervening assistance system that could prevent truck to cyclist accidents. The main challenges are local sight obstructions such as fences, hedges, etc., rule violations by cyclists, and the complexity of correctly predicting the cyclist’s intentions, i.e., detecting the trajectory. Taking these accident circumstances into consideration, a BSD system could prevent between 26.3% and 65.8% of accidents involving HGVs and cyclists. Full article