Crowd Density Research Articles

Exploring the Impacts of Exit Structures on Evacuation Efficiency

In the context of a fire emergency, safe and efficient exits are of paramount importance for pedestrian evacuation. The recent rapid development in the construction industry has rendered exit structures more diverse and complex. However, little attention has been paid to the influence of exit structures on the efficiency of crowd evacuation processes. In this paper, a tentative experiment was designed to preliminarily reveal the effects of five exit structures (Exit 1, Exit 2, Exit 4, Exit 5, and Exit 3 as examples for comparison) on crowd evacuation. Exit 1 has door leaves opening outward. Exit 2 has door leaves opening inward. Exit 3 has no leaves. Exit 4 consists of double-layer exit doors with the doors opening outward. Exit 5 comprises double-layer exit doors with the doors opening both sides outwards Subsequently, according to the properties of this experiment, a social force-based simulation model was established using the AnyLogic software 8.8.4. By changing the exit width and the crowd density, data such as evacuation time, flow rate, crowd density, and time delay were investigated in detail. The results revealed a notable variation in the evacuation efficiency depending on the deign of the exit. The respective flow rates for Exits 1, 2, 3, 4, and 5 were 0.66 people/(m·s), 0.77 people/(m·s), 0.80 people/(m·s), 0.71 people/(m·s), and 0.66 people/(m·s). Although Exit 3 excelled in terms of evacuation efficiency, it is not directly applicable to real architectural structures. Therefore, Exit 2 emerged as a highly promising solution in terms of flow rate and population control in the exit area, underscoring the effectiveness and practicality of its structural design. It is prospective that the results of this study can offer engineering and technical professionals valuable references and guidance concerning the design of exit structures.

Public Place Crowd Transaction Monitoring System

Currently, the phenomenon of abnormal movement in public spaces by groups is becoming increasingly prominent, leading to issues concerning public flow and safety. The escalating problems of high crowd density, the presence of controlled dangerous items, and unexpected group activities highlight the necessity for timely detection in public settings. Timely identification of such scenarios will facilitate prompt responses and assistance from relevant government departments. Exploring how artificial intelligence technology can aid urban management personnel in effectively detecting abnormal group behaviors is crucial. Having the ability to swiftly and efficiently evacuate crowds in emergency situations holds significant practical importance. This paper employs deep learning methodologies to assist urban management personnel in efficiently monitoring crowd density and detecting abnormal behaviors. The aim is to maintain crowd density within reasonable limits and enable rapid and effective crowd evacuation in emergency situations. Detection of abnormal group behaviors typically involves methods based on global features, extracting feature patterns like optical flow from entire video segments and constructing corresponding histograms. Given that automatic classification of crowd patterns involves sudden and abnormal changes, a novel method is proposed to extract motion "textures" from dynamic STV (Space-Time Volume) blocks formed from real-time video streams.

An extended social force model on unidirectional flow considering psychological and behavioral impacts of hazard source

An accurate assessment of the evacuation efficiency in case of disasters is of vital importance to the safety design of buildings and street blocks. Hazard sources not only physically but psychologically affect the pedestrians, which may further alter their behavioral patterns. This effect is especially significant in narrow spaces, such as corridors and alleys. This study aims to integrate a non-spreading hazard source into the social force model following the results from a previous experiment and simulation, and to simulate unidirectional pedestrian flows over various crowd densities and clarity–intensity properties of the hazard source. The integration include a virtual repulsion force from the hazard source and a decay on the social force term. The simulations reveal (i) that the hazard source creates virtual bottlenecks that suppress the flow, (ii) that the inter-pedestrian push forms a stabilisation phase on the flow-density curve within medium-to-high densities, and (iii) that the pedestrians are prone to a less orderly and stable pattern of movement in low clarity–intensity scenarios, possibly with lateral collisions passing the hazard source.

CAGNet: Coordinated attention guidance network for RGB-T crowd counting

Estimating crowd density is a demanding task that has garnered significant research attention in urban planning, intelligent transportation, and other related fields. This study utilizes RGB and thermal images to leverage multimodal information and introduces a coordinated attention guidance network (CAGNet) for RGB-thermal (RGB-T) crowd counting. The framework enhances the expressive capabilities of backbone features by incorporating overall and local relationships through the information enhancement unit module, which utilizes the context coordination perception module for horizontal information mining, interactively compensating for the continuity of spatial information. Subsequently, it utilizes diverse multi-level information features for hierarchical intersection and progression, resulting in an accurate crowd counting density map. Experimental results on the RGBT-CC benchmark dataset demonstrate the robustness and effectiveness of CAGNet for RGB-T crowd counting. Furthermore, the proposed CAGNet can be extended to crowd density estimation and has achieved high performance on the ShanghaiTechRGBD and DroneRGBT datasets. The former dataset comprises paired RGB images and depth maps. The code and model for this research are available at https://github.com/WBangG/CAGNet.

Psychological pushing propagation in crowds—Does the observation of pushing behavior promote further intentional pushing?

When large numbers of people come together (e.g., at concerts or religious gatherings), critical situations can arise easily. While physical factors such as crowd density play a role, people's behavior can also affect crowd dynamics. For example, pushing and shoving, which are closely related to density, can quickly contribute to potentially dangerous dynamics. There is little extant research, however, on why people start pushing in the first place. Aside from individual reasons (e.g., motivation), social reasons might also play a role: an initial instance of pushing might be imitated or spark a competition if the pusher seems to reach the goal faster or the behavior of individuals defines a group norm whether pushing is allowed or not. Practically speaking, these social factors should lead people to push because they perceive other pushers, or, in other words, a psychological pushing propagation occurs. To address this question, the behavior (pushing or non-pushing) of people in 14 different experimental runs of crowds walking through a bottleneck (N = 776) was assessed by two independent raters with the help of a rating system of forward motion. This assessment was then linked to the spatiotemporal positions of the participants to combine it with the neighborhood relations. Based on that, it was analyzed whether individuals who started to push were more likely to be in the direct neighborhood of individuals who were already pushing. Results showed a small but significant effect suggesting that there is an influence, but that pushing is not overly “contagious.”

Risk-Aware Deep Reinforcement Learning for Robot Crowd Navigation

Ensuring safe and efficient navigation in crowded environments is a critical goal for assistive robots. Recent studies have emphasized the potential of deep reinforcement learning techniques to enhance robots’ navigation capabilities in the presence of crowds. However, current deep reinforcement learning methods often face the challenge of robots freezing as crowd density increases. To address this issue, a novel risk-aware deep reinforcement learning approach is proposed in this paper. The proposed method integrates a risk function to assess the probability of collision between the robot and pedestrians, enabling the robot to proactively prioritize pedestrians with a higher risk of collision. Furthermore, the model dynamically adjusts the fusion strategy of learning-based and risk-aware-based features, thereby improving the robustness of robot navigation. Evaluations were conducted to determine the effectiveness of the proposed method in both low- and high-crowd density settings. The results exhibited remarkable navigation success rates of 98.0% and 93.2% in environments with 10 and 20 pedestrians, respectively. These findings emphasize the robust performance of the proposed method in successfully navigating through crowded spaces. Additionally, the approach achieves navigation times comparable to those of state-of-the-art methods, confirming its efficiency in accomplishing navigation tasks. The generalization capability of the method was also rigorously assessed by subjecting it to testing in crowd environments exceeding the training density. Notably, the proposed method attains an impressive navigation success rate of 90.0% in 25-person environments, surpassing the performance of existing approaches and establishing itself as a state-of-the-art solution. This result highlights the versatility and effectiveness of the proposed method in adapting to various crowd densities and further reinforces its applicability in real-world scenarios.

Cloud‐fog assisted human tracing and disaster evacuation framework

AbstractMany kinds of natural hazards and man‐made calamities have occurred in recent years, each of which has left a huge number of victims and has been challenging to deal with. The solution is an effective evacuation planning system, which is gaining the interest of governments, academic institutions, and industries all across the world. This study focuses on quick and organized evacuation from densely populated areas during disasters to decrease damage successfully. This research proposes a fog‐cloud‐assisted effective disaster evacuation framework. The data accumulation layer uses IoT, mobile networks, and GPS technologies to collect environmental and location‐based data to track the inhabitants. The fog layer is utilized for (i) disaster event detection and (ii) human tracing and crowd density analysis. A cloud layer makes it easier for an evacuation algorithm to produce an evacuation map that will point evacuees to the exit while computing a quick and safe path utilizing environmental and inhabitant information gathered from the fog. The findings reveal that the performance of Model A surpasses the split delivery vehicle routing problem model with the minimum demand of vehicles and with improvement of 38.1% average. Also, A* takes less time (18.3 s) than other state‐of‐the‐art techniques to find the best optimal route for evacuation. Implementation and performance analysis demonstrate the efficiency and effectiveness of the proposed framework.

Simulation of bi-directional pedestrian flow in corridor based on direction fuzzy visual field

Bi-directional pedestrian flow in corridors is a complex dynamic system due to the diversity in pedestrian psychological characteristics. Incorporating individual differences of pedestrians is vital for improving pedestrian flow models. However, due to the inherent complexity and variability of pedestrian movement, model parameter calibration remains challenging. Controlled experiments are needed to collect empirical pedestrian movement data under different environments. This enriches the database on pedestrian movement patterns and provides necessary support for improving pedestrian flow models. To address this issue, we conducted controlled experiments to quantify pedestrian heterogeneity by defining the direction of fuzzy visual field (DFVF). The DFVF incorporates various static and dynamic pedestrian factors. We used it to modify the traditional cellular automata model. This improved model simulates bi-directional pedestrian movements in the corridors, reproduces density-speed and density-volume relationships, and reveals self-organization phenomena. Furthermore, an analysis was conducted to examine the impacts of pedestrian density and facility spatial layout on evacuation time. Pedestrian interactions were also studied to uncover fundamental bi-directional flow properties. As pedestrian density increased, the evacuation time showed an exponential upward trend. Corridor length significantly impacts evacuation time, while increasing corridor width helps control it. As crowd density increases, pedestrian flows exhibit three distinct steady states: the strolling flow at low densities, directional separated flows at medium densities, and dynamic multi-lane flows at high densities. In summary, the modified cellular automata model successfully incorporates pedestrian heterogeneity and reveals intrinsic bi-directional pedestrian flow patterns. This study provides valuable insights for pedestrian facility design and optimizing pedestrian flow organization.

Crowd Density Estimation Method based on a Cascaded Multilevel Convolution Neural Network

Crowd Density Estimation Method based on a Cascaded Multilevel Convolution Neural Network

Multi-Scene Mask Detection Based on Multi-Scale Residual and Complementary Attention Mechanism

Vast amounts of monitoring data can be obtained through various optical sensors, and mask detection based on deep learning integrates neural science into a variety of applications in everyday life. However, mask detection poses technical challenges such as small targets, complex scenes, and occlusions, which necessitate high accuracy and robustness in multi-scene target detection networks. Considering that multi-scale features can increase the receptive field and attention mechanism can improve the detection effect of small targets, we propose the YOLO-MSM network based on the multi-scale residual (MSR) block, multi-scale residual cascaded channel-spatial attention (MSR-CCSA) block, enhanced residual CCSA (ER-CCSA) block, and enhanced residual PCSA (ER-PCSA) block. Considering the performance and parameters, we use YOLOv5 as the baseline network. Firstly, for the MSR block, we construct hierarchical residual connections in the residual blocks to extract multi-scale features and obtain finer features. Secondly, to realize the joint attention function of channel and space, both the CCSA block and PCSA block are adopted. In addition, we construct a new dataset named Multi-Scene-Mask, which contains various scenes, crowd densities, and mask types. Experiments on the dataset show that YOLO-MSM achieves an average precision of 97.51%, showing better performance than other detection networks. Compared with the baseline network, the mAP value of YOLO-MSM is increased by 3.46%. Moreover, we propose a module generalization improvement strategy (GIS) by training YOLO-MSM on the dataset augmented with white Gaussian addition noise to improve the generalization ability of the network. The test results verify that GIS can greatly improve the generalization of the network and YOLO-MSM has stronger generalization ability than the baseline.

STP4: spatio temporal path planning based on pedestrian trajectory prediction in dense crowds

This article proposes a means of autonomous mobile robot navigation in dense crowds based on predicting pedestrians’ future trajectories. The method includes a pedestrian trajectory prediction for a running mobile robot and spatiotemporal path planning for when the path crosses with pedestrians. The predicted trajectories are converted into a time series of cost maps, and the robot achieves smooth navigation without dodging to the right or left in crowds; the path planner does not require a long-term prediction. The results of an evaluation implementing this method in a real robot in a science museum show that the trajectory prediction works. Moreover, the proposed planning’s arrival times is 26.4% faster than conventional 2D path planning’s arrival time in a simulation of navigation in a crowd of 50 people.

Optimizing University Mobility : An Internal Navigation and Crowd Management System

In the evolving landscape of educational technology, the article explores the critical frontier of indoor navigation systems, focusing on universities. Traditional approaches in higher education often fall short of meeting dynamic user expectations, necessitating revolutionary solutions. This research introduces an innovative internal navigation and crowd management system that seamlessly integrates augmented reality, natural language processing, machine learning, and image processing technologies. The Android platform serves as the foundation, harnessing augmented reality's transformative capabilities to provide real-time visual cues and personalized wayfinding experiences. The voice interaction module, backed by NLP and ML, creates an intelligent, context-aware assistant. The crowd management module, employing advanced image processing, delivers real-time crowd density insights. Personalized recommendations, powered by NLP and ML, offer tailored canteen suggestions based on user preferences. The agmented reality navigation module, using Mapbox, Unity Hub, AR Core, and Vuforia, enriches the user experience with dynamic visual cues. Results reveal the success of each module: the voice interaction module showcases continuous learning, user-centric feedback, contextual guidance excellence, robust security, and multimodal interaction flexibility. The crowd management module excels in video feed processing, image processing with OpenCV, and real-time availability information retrieval. The personalized recommendations module demonstrates high accuracy, equilibrium, and robust performance. The AR navigation module impresses with precision, enriched navigation, and tailored routes through machine learning. This cohesive system sets new benchmarks for user-centric technology in universities. Future work includes multi-university integration, intelligent spatial design, and real-time decision support, paving the way for more efficient, user-centered university experiences and contributing to the advancement of smart university environments. The research serves as a pivotal force in reshaping interactions within university spaces, envisioning a future where technology seamlessly enhances the essence of human interaction in educational environments.

A Fusion-Based Dense Crowd Counting Method for Multi-Imaging Systems

Dense crowd counting has become an essential technology for urban security management. The traditional crowd counting methods mainly apply to the scene with a single view and obvious features but cannot solve the problem with a large area and fuzzy crowd features. Therefore, this paper proposes a crowd counting method based on high and low view information fusion (HLIF) for large and complex scenes. First, a neural network based on an attention mechanism (AMNet) is established to obtain a global density map from a high view and crowd counts from a low view. Then, the temporal correlation and spatial complementarity between cameras are used to calibrate the overlap areas of the two images. Finally, the total number of people is calculated by combining the low-view crowd counts and the high-view density map. Compared to single-view crowd counting methods, HLIF is experimentally more accurate and has been successfully applied in practice.

A lightweight dense crowd density estimation network for efficient compression models

Crowd density estimation is a task of intelligent applications, and its operation efficiency is very important. However, to obtain a better density estimation performance, most of the existing works often design larger and more complex network structures, which will result in them occupying considerable memory, time and other resources at runtime, and require the support of high-performance hardware platforms, which are difficult to apply in practice. In this paper, to overcome the above problems, we propose a lightweight dense crowd estimation method based on channel attention multi-scale feature fusion. Specifically, in the process of feature extraction, an efficient and lightweight convolution module (L-weight) is designed to extract crowd features in stages, which reduces the amount of network parameters and computing costs, and we capture multi-scale crowd information through the feature extraction network of pyramid structure, which solves the problem of uneven crowd scale in video images. In the process of feature fusion, a channel attention fusion module is designed, which weights and fuses the feature information of different scales, effectively fuses multi-scale information and suppresses useless information. In addition, we design a new loss function, which enhances the sensitivity of the crowd through the pixel space loss (L2), counting loss (LC) and structural similarity loss (LS), to ensure the counting accuracy. Extensive experiments on four mainstream datasets demonstrate that compared with other state-of-the-art methods, our method achieves an optimal trade-off between counting performance and running speed, and is suitable for low-performance computing platforms such as embedded.

SafeCrowdNav: safety evaluation of robot crowd navigation in complex scenes.

Navigating safely and efficiently in dense crowds remains a challenging problem for mobile robots. The interaction mechanisms involved in collision avoidance require robots to exhibit active and foresighted behaviors while understanding the crowd dynamics. Deep reinforcement learning methods have shown superior performance compared to model-based approaches. However, existing methods lack an intuitive and quantitative safety evaluation for agents, and they may potentially trap agents in local optima during training, hindering their ability to learn optimal strategies. In addition, sparse reward problems further compound these limitations. To address these challenges, we propose SafeCrowdNav, a comprehensive crowd navigation algorithm that emphasizes obstacle avoidance in complex environments. Our approach incorporates a safety evaluation function to quantitatively assess the current safety score and an intrinsic exploration reward to balance exploration and exploitation based on scene constraints. By combining prioritized experience replay and hindsight experience replay techniques, our model effectively learns the optimal navigation policy in crowded environments. Experimental outcomes reveal that our approach enables robots to improve crowd comprehension during navigation, resulting in reduced collision probabilities and shorter navigation times compared to state-of-the-art algorithms. Our code is available at https://github.com/Janet-xujing-1216/SafeCrowdNav.

Foreground Segmentation-Based Density Grading Networks for Crowd Counting.

Estimating object counts within a single image or video frame represents a challenging yet pivotal task in the field of computer vision. Its increasing significance arises from its versatile applications across various domains, including public safety and urban planning. Among the various object counting tasks, crowd counting is particularly notable for its critical role in social security and urban planning. However, intricate backgrounds in images often lead to misidentifications, wherein the complex background is mistaken as the foreground, thereby inflating forecasting errors. Additionally, the uneven distribution of crowd density within the foreground further exacerbates predictive errors of the network. This paper introduces a novel architecture with a three-branch structure aimed at synergistically incorporating hierarchical foreground information and global scale information into density map estimation, thereby achieving more precise counting results. Hierarchical foreground information guides the network to perform distinct operations on regions with varying densities, while global scale information evaluates the overall density level of the image and adjusts the model's global predictions accordingly. We also systematically investigate and compare three potential locations for integrating hierarchical foreground information into the density estimation network, ultimately determining the most effective placement.Through extensive comparative experiments across three datasets, we demonstrate the superior performance of our proposed method.

Crowd counting algorithm based on face detection and skin color recognition

This paper introduces an innovative crowd counting algorithm using skin color information. Through stages of color space transformation, threshold segmentation, morphological processing, and region filtering, the algorithm successfully conducts crowd counting in images. The study encompasses analyses of images with diverse crowd densities, skin colors, backgrounds, and lighting intensities, revealing the algorithm's robustness to various factors. It remains unaffected by skin color and crowd size and exhibits minimal sensitivity to background and lighting intensity. Furthermore, the paper explores image feature analysis and uses MATLAB programming for simulation and initial crowd counting, considering images with different actual crowd sizes. Despite minor issues such as the insufficient separation of faces from clothing and the influence of lighting intensity, the algorithm performs reliably in most scenarios, demonstrating high crowd counting accuracies. To bolster the accuracy and robustness of the algorithm, optimization of the separation step and control of the lighting effect on images is suggested. The key focus of this study is the application of the Gaussian model in the YCbCr color space for face detection and examining its impact on the efficiency and accuracy of crowd counting algorithm. The research not only provides a novel approach for crowd counting in images but also offers insightful perspectives for future studies and potential improvements. Thus, the study proves to be a significant contribution to face detection and recognition technology, enhancing its application in fields like public safety, crowd management, and surveillance systems.

Influence of walkway slope on single-file pedestrian flow dynamics: Results from an experimental study

Sloping walkways are common pedestrian traffic facilities in modern cities, especially, in some cities located in mountainous terrain, such as Hong Kong, and Zurich. Compared with level walkways, sloping walkways may lead to the decrease in pedestrian flow efficiency, and are more prone to trigger crowd disasters at extreme crowd densities, for example, recent crowd stampede event in Itaewon, South Korea, did happen on a narrow ramp with an extreme crowd density. Therefore, it is important to study pedestrian flow dynamics on ramps. In this paper, we conducted single-file pedestrian flow experiments in 0°, 3°, 5°, 7°, 9°, 12°, 17°, 22°, 27° uphill and downhill conditions, and analyzed contrastively the influence of walkway slope on pedestrian flow dynamics. Our main findings are as follows. (i) Pedestrian flow efficiency is the highest in 7° downhill condition, indicating that a small downhill slope is advantageous to pedestrian flowing. (ii) Whether it is uphill or downhill, the ramp steeper than 7° will significantly decrease pedestrian flow efficiency, therefore is disadvantageous to pedestrian flowing. (iii) For a ramp, uphill and downhill pedestrian flow efficiencies are not equal. Downhill pedestrian flow efficiency is always larger than uphill pedestrian flow efficiency. These findings provide important guidance for the design of sloping walkways and the safety management of pedestrian crowds on sloping walkways.

Dual-branch counting method for dense crowd based on self-attention mechanism

A dense crowd counting method based on self-attention mechanism with dual-branch fusion network is proposed in this paper. Our method aims to address the problems of large variations in head scales and complex backgrounds in dense crowd images. This method combines the CNN and Transformer network frameworks and consists of shallow feature extraction network, dual-branch fusion network, and deep feature extraction network. The VGG16 network is employed by the shallow feature extraction network to extract low-level features. A multi-scale CNN branch and a Transformer branch built on an improved self-attention module make up the dual-branch fusion network, which collects local and global information on crowd areas, respectively. The Transformer network, which is based on a mixed attention module, is employed by the deep feature extraction network to further separate complicated backgrounds and concentrate on crowd areas. Both counting-level weakly supervised and location-level fully supervised methods are employed in the experiments. On four widely used datasets, the results demonstrate that the proposed method outperforms the most recent research. Our method has a higher counting accuracy with low parameter volumes and a counting accuracy of 89.1% under full supervision when compared to existing weakly supervised methods. The results of the experiments demonstrate that the method has excellent crowd counting performance and can accurately count in high-density and high-occlusion scenes.

The Group Interaction Field for Learning and Explaining Pedestrian Anticipation

Anticipating others’ actions is innate and essential in order for humans to navigate and interact well with others in dense crowds. This ability is urgently required for unmanned systems such as service robots and self-driving cars. However, existing solutions struggle to predict pedestrian anticipation accurately, because the influence of group-related social behaviors has not been well considered. While group relationships and group interactions are ubiquitous and significantly influence pedestrian anticipation, their influence is diverse and subtle, making it difficult to explicitly quantify. Here, we propose the group interaction field (GIF), a novel group-aware representation that quantifies pedestrian anticipation into a probability field of pedestrians’ future locations and attention orientations. An end-to-end neural network, GIFNet, is tailored to estimate the GIF from explicit multidimensional observations. GIFNet quantifies the influence of group behaviors by formulating a group interaction graph with propagation and graph attention that is adaptive to the group size and dynamic interaction states. The experimental results show that the GIF effectively represents the change in pedestrians’ anticipation under the prominent impact of group behaviors and accurately predicts pedestrians’ future states. Moreover, the GIF contributes to explaining various predictions of pedestrians’ behavior in different social states. The proposed GIF will eventually be able to allow unmanned systems to work in a human-like manner and comply with social norms, thereby promoting harmonious human–machine relationships.