• Journal of IKEEE
• /
• v.17 no.4
• /
• pp.469-477
• /
• 2013

This paper enhances the speed of a pedestrian detection using an estimation of feature information based on integral image. Pedestrian model or input image should be resized to the size of various pedestrians. In case that the size of pedestrian model would be changed, pedestrian models with respect to the size of pedestrians should be required. Reducing the size of pedestrian model, however, deteriorates the quality of the model information. Since various features according to the size of pedestrian models should be extracted, repetitive feature extractions spend the most time in overall process of pedestrian detection. In order to enhance the processing time of feature extraction, this paper proposes the fast extraction of pedestrian features based on the estimate of integral image. The efficiency of the proposed method is evaluated by comparative experiments with the Channel Feature and Adaboost training using INRIA person dataset.

• Journal of Korean Society of Transportation
• /
• v.27 no.1
• /
• pp.169-177
• /
• 2009

In this paper, the rotated hexagonal lattice model (RHLM) was proposed, which is applied to pedestrian flow, and developed the simulation model for the pedestrian counterflow. RHLM is an upgrade version of the square lattice model(SLM) and hexagonal lattice model(HLM). The simulation was performed at the hexagonal lattice $20{\times}20$ and evaluated by different speed, density and flow conditions. Simulation results are compared with SLM and show that RHLM can replicate the characteristics of pedestrian traffic more effectively and reliably than any other existing models from several perspectives. First, RHLM can explain the shortest-path movement of pedestrians and more realistic avoidance motion. If they cannot move straight direction, they can move shorter distance from previous position to destination. Second, RHLM reflects the characteristics that the pedestrian can move with higher capacity and the speed of pedestrian flow is hard to zero.

• Journal of Information Processing Systems
• /
• v.16 no.6
• /
• pp.1309-1323
• /
• 2020

Crowd evacuation simulation is an important research issue for designing reasonable building layouts and planning more effective evacuation routes. The social force model (SFM) is an important pedestrian movement model, and is widely used in crowd evacuation simulations. The model can effectively simulate crowd evacuation behaviors in a simple scene, but for a multi-obstacle scene, the model could result in some undesirable problems, such as pedestrian evacuation trajectory oscillation, pedestrian stagnation and poor evacuation routing. This paper analyzes the causes of these problems and proposes an improved SFM for complex multi-obstacle scenes. The new model adds navigation points and walking shortest route principles to the SFM. Based on the proposed model, a crowd evacuation simulation system is developed, and the crowd evacuation simulation was carried out in various scenes, including some with simple obstacles, as well as those with multi-obstacles. Experiments show that the pedestrians in the proposed model can effectively bypass obstacles and plan reasonable evacuation routes.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.18 no.3
• /
• pp.104-109
• /
• 2010

The fatality of pedestrian accounts for about 21.2% of all fatality at 2007 year in Korea. In car to pedestrian accident it is very important to inspect the throw distance of pedestrian after collision for exact reconstructing of the accident. The variables that influence on the throw distance of pedestrian can be classified into the factors of vehicle and pedestrian, and road condition. It was simulated by PC-CRASH, a kinetic analysis program for a traffic accident in sedan type vehicle and SPSS program was used for regression analysis. From the results, the throw distance of pedestrian increased with the increasing of vehicle velocity, and decreased with the increasing of impact offset. Also it decreased with the increasing of velocity of pedestrian at accident, and throw distance at the road condition of wet was longer than that at dry condition. Finally, the regression model of sedan type vehicle on the throw distance of pedestrian was as follows; $$dist_i=2.39-0.11offset_i+0.59speed_i-545height_i-0.25walk_i+2.78wet_i+{\epsilon}_i$$.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.12
• /
• pp.3904-3922
• /
• 2022

As a research hotspot, pedestrian detection has a wide range of applications in the field of computer vision in recent years. However, current pedestrian detection methods have problems such as insufficient detection accuracy and large models that are not suitable for large-scale deployment. In view of these problems mentioned above, a lightweight pedestrian detection and early warning method using a new model called you only look once (Yolov5) is proposed in this paper, which utilizing advantages of Yolov5s model to achieve accurate and fast pedestrian recognition. In addition, this paper also optimizes the loss function of the batch normalization (BN) layer. After sparsification, pruning and fine-tuning, got a lot of optimization, the size of the model on the edge of the computing power is lower equipment can be deployed. Finally, from the experimental data presented in this paper, under the training of the road pedestrian dataset that we collected and processed independently, the Yolov5s model has certain advantages in terms of precision and other indicators compared with traditional single shot multiBox detector (SSD) model and fast region-convolutional neural network (Fast R-CNN) model. After pruning and lightweight, the size of training model is greatly reduced without a significant reduction in accuracy, and the final precision reaches 87%, while the model size is reduced to 7,723 KB.

• Journal of the Korean Society of Safety
• /
• v.29 no.3
• /
• pp.98-106
• /
• 2014

This study proposes severity analysis for pedestrian accidents by improving variables which were used for general severity analysis. The existing variables were collected based on the interviews with policeman or witnesses and evidence of accidents. Therefore, existing variables were subjective and had several measurement errors. In order to improve such problems, this study collected variables from vehicle recorder of taxi which recorded the moment of accidents. As a result, explanatory power of independent variables was enhanced and the complete objective variables could be collected. After collecting variables, ordered probit model was developed by utilizing vehicle recorder database. Fitness of ordered probit model was 0.23. Vehicle speed and pedestrian's eye direction variables were the most critical factors for severity of pedestrian accident. In addition, severity analysis for vulnerable pedestrian was carried out. As a result, it was revealed that vehicle speed, pedestrian's eye direction and safety zone variables affected the severity of pedestrian accidents most. Particularly, vehicle speed variable is the most important factor. Consequently, driver's defensive driving and compliance to the regulations are the priority to reduce severity of pedestrian accidents and prevent pedestrian accident.

• Journal of the Architectural Institute of Korea Planning & Design
• /
• v.34 no.11
• /
• pp.23-32
• /
• 2018

The purpose of this study is to establish a walking energy weighted ERAM model that can predict the pedestrian volume by the connection structure of the vertical and horizontal spaces within a three-dimensional building. The process of building a walking-energy weighted ERAM model is as follows. First, the spatial graph was used to reproduce three-dimensional buildings with vertical and horizontal spatial connection structures. Second, the walking energy was measured on the spatial graph. Third, ERAM model was used to apply weights with spatial connection properties in random walking environment, and the walking energy weights were applied to the ERAM model to calculate the walk energy weighted ERAM values and visualize the distribution of pedestrian flow. To verify the validation of the established model, existing and proposed spatial analysis models were compared to real space. The results of this study are as follows : The model proposed in this study showed as much elaborated estimation of pedestrian traffic flow in real space as in traditional spatial analysis models, and also it showed much higher level of forecasting pedestrian traffic flow in real space than existing models.

• Journal of Power System Engineering
• /
• v.13 no.2
• /
• pp.56-62
• /
• 2009

The fatalities of pedestrian account for about 40.0% of all fatalities in Korea 2005. Vehicle-Pedestrian accident generates trajectory of pedestrian. In pedestrian involved accident, the most important data to inspect accident is throw distance of pedestrian. The throw distance of pedestrian can be influenced by many variables. But existing studies have been done for simple factors. The variables that influence trajectory of pedestrian can be classified into vehicular factors, pedestrian factors, and road factors. The trajectory of pedestrian, dynamic characteristics of multi-body were analyzed by PC-CRASH, a kinetic analysis program for a traffic accident. PC-CRASH enables an analyst to investigate the effect of many variables. The influence of the offset of impact point was analyzed by Working Model. Based on the results, the variables that influence trajectory of pedestrian were vehicular frontal shape, vehicular impact speed, the offset of impact point, the height of pedestrian, friction coefficients of pedestrian. However the weight of pedestrian did not affect trajectory of pedestrian considerably.

• Journal of the Korean Society for Precision Engineering
• /
• v.24 no.5
• /
• pp.110-117
• /
• 2007

According to the pedestrian protection regulations of Europe and Japan, the head injury must not exceed a limitation in the defined test condition for the protection of pedestrians from a vehicle crash. However, it is difficult to evaluate the performance of protection because each regulation has different test conditions such as dummy, impact speed and so on. This circumstance needs the development of a model that describes the anthropometry of the crash victim with a sufficient accuracy. We constructed scaled pedestrian dummies using MADYSCALE. Simulations were performed for various crash speeds and pedestrian postures. The scaled Korean dummies and HybridIII dummies were used to compare the pedestrian dynamic behaviors and head injury criteria during the collision. The HIC values of scaled korean dummies were found to be higher than those of Hybrid III dummies. The impact for gait posture was less than that for standing.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.6 no.3
• /
• pp.71-76
• /
• 2007

The fatalities of pedestrian account for about 40.0% of all fatalities in Korea 2005. Vehicle-Pedestrian accident generates trajectory of pedestrian. In pedestrian involved accident, the most important data to inspect accident is throw distance of pedestrian. The throw distance of pedestrian can be influenced by many variables. The variables that influence trajectory of pedestrian can be classified into vehicular factors, pedestrian factors, and road factors. Vehicular factors are the frontal shape of vehicle, impact speed of vehicle, the offset of impact point. Many studies have been done about the relation between impact speed and throw distance of pedestrian. But the influence of the offset of impact point was neglected. The influence of the offset of impact point was analyzed by Working Model, and the trajectory of pedestrian, dynamic characteristics of multi-body were analyzed by PC-CRASH, a kinetic analysis program for a traffic accident. Based on the results, the increase of offset reduced the throw distance of pedestrian. However box type vehicle just like bus, the offset of impact point did not influence the throw distance of pedestrian considerably.