• International Journal of Internet, Broadcasting and Communication
• /
• 제12권4호
• /
• pp.83-92
• /
• 2020

A recent study to determine the fall is focused on analyzing fall motions using a recurrent neural network (RNN), and uses a deep learning approach to get good results for detecting human poses in 2D from a mono color image. In this paper, we investigated the improved detection method to estimate the position of the head and shoulder key points and the acceleration of position change using the skeletal key points information extracted using PoseNet from the image obtained from the 2D RGB low-cost camera, and to increase the accuracy of the fall judgment. In particular, we propose a fall detection method based on the characteristics of post-fall posture in the fall motion analysis method and on the velocity of human body skeleton key points change as well as the ratio change of body bounding box's width and height. The public data set was used to extract human skeletal features and to train deep learning, GRU, and as a result of an experiment to find a feature extraction method that can achieve high classification accuracy, the proposed method showed a 99.8% success rate in detecting falls more effectively than the conventional primitive skeletal data use method.

• 국제학술발표논문집
• /
• The 9th International Conference on Construction Engineering and Project Management
• /
• pp.361-369
• /
• 2022

Unsafe actions and behaviors of workers cause most accidents at construction sites. Nowadays, occupational safety is a top priority at construction sites. However, this problem often requires money and effort from investors or construction owners. Therefore, decreasing the accidents rates of workers and saving monitoring costs for contractors is necessary at construction sites. This study proposes an unsafe behavior detection method based on a skeleton model to classify three common unsafe behaviors on the scaffold: climbing, jumping, and running. First, the OpenPose method is used to obtain the workers' key points. Second, all skeleton datasets are aggregated from the temporary size. Third, the key point dataset becomes the input of the action classification model. The method is effective, with an accuracy rate of 89.6% precision and 90.5% recall of unsafe actions correctly detected in the experiment.

• Structural Engineering and Mechanics
• /
• 제83권3호
• /
• pp.305-326
• /
• 2022

To realize the recycling utilization of waste concrete and alleviate the shortage of resources, 11 specimens of steel reinforced recycled concrete (SRRC) filled circular steel tube columns were designed and manufactured in this study, and the cyclic loading tests on the specimens of columns were also carried out respectively. The hysteretic curves, skeleton curves and performance indicators of columns were obtained and analysed in detail. Besides, the finite element model of columns was established through OpenSees software, which considered the adverse effect of recycled coarse aggregate (RA) replacement rates and the constraint effect of circular steel tube on internal RAC. The numerical calculation curves of columns are in good agreement with the experimental curves, which shows that the numerical model is relatively reasonable. On this basis, a series of nonlinear parameters analysis on the hysteretic behaviors of columns were also investigated. The results are as follows: When the replacement rates of RA increases from 0 to 100%, the peak loads of columns decreases by 7.78% and the ductility decreases slightly. With the increase of axial compression ratio, the bearing capacity of columns increases first and then decreases, but the ductility of columns decreases rapidly. Increasing the wall thickness of circular steel tube is very profitable to improve the bearing capacity and ductility of columns. When the section steel ratio increases from 5.54% to 9.99%, although the bearing capacity of columns is improved, it has no obvious contribution to improve the ductility of columns. With the decrease of shear span ratio, the bearing capacity of columns increases obviously, but the ductility decreases, and the failure mode of columns develops into brittle shear failure. Therefore, in the engineering design of columns, the situation of small shear span ratio (i.e., short columns) should be avoided as far as possible. Based on this, the calculation model on the skeleton curves of columns was established by the theoretical analysis and fitting method, so as to determine the main characteristic points in the model. The effectiveness of skeleton curve model is verified by comparing with the test skeleton curves.

• Earthquakes and Structures
• /
• 제25권5호
• /
• pp.385-398
• /
• 2023

The parameters of the multiple-vertical-line-element model (MVLEM) of reinforced concrete (RC) shear walls are often empirically determined, which causes large simulation errors. To improve the simulation accuracy of the MVLEM for RC shear walls, this paper proposed a novel method to determine the MVLEM parameters using the artificial neural network (ANN). First, a comprehensive database containing 193 shear wall specimens with complete parameter information was established. And the shear walls were simulated using the classic MVLEM. The average simulation errors of the lateral force and drift of the peak and ultimate points on the skeleton curves were approximately 18%. Second, the MVLEM parameters were manually optimized to minimize the simulation error and the optimal MVLEM parameters were used as the label data of the training of the ANN. Then, the trained ANN was used to generate the MVLEM parameters of the collected shear walls. The results show that the simulation error of the predicted MVLEM was reduced to less than 13% from the original 18%. Particularly, the responses generated by the predicted MVLEM are more identical to the experimental results for the testing set, which contains both flexure-control and shear-control shear wall specimens. It indicates that establishing MVLEM for RC shear walls using ANN is feasible and promising, and that the predicted MVLEM substantially improves the simulation accuracy.

• Structural Engineering and Mechanics
• /
• 제72권3호
• /
• pp.355-366
• /
• 2019

Self-centering wall (SCW) is a resilient and sustainable structural system which incorporates unbonded posttensioning (PT) tendons to provide self-centering (SC) capacity along with supplementary dissipators to dissipate seismic energy. Hysteretic energy dissipators are usually placed at two sides of SCWs to facilitate ease of postearthquake examination and convenient replacement. To achieve a good prediction for the skeleton curve of the wall, this paper firstly developed an analytical investigation on lateral load responses of self-centering walls with distributed vertical dampers (VD-SCWs) using the concept of elastic theory. A simplified method for the calculation of limit state points is developed and validated by experimental results and can be used in the design of the system. Based on the analytical results, parametric analysis is conducted to investigate the influence of damper and tendon parameters on the performance of VD-SCWs. The results show that the proposed approach has a better prediction accuracy with less computational effects than the Perez method. As compared with previous experimental results, the proposed method achieves up to 60.1% additional accuracy at the effective linear limit (DLL) of SCWs. The base shear at point DLL is increased by 62.5% when the damper force is increased from 0kN to 80kN. The wall stiffness after point ELL is reduced by 69.5% when the tendon stiffness is reduced by 75.0%. The roof deformation at point LLP is reduced by 74.1% when the initial tendon stress is increased from $0.45f_{pu}$ to $0.65f_{pu}$.

• 한국산업융합학회 논문집
• /
• 제24권6_1호
• /
• pp.693-698
• /
• 2021

As technology advances, there is increasing need for research in different fields where this technology is applied. On of the most researched topic in computer vision is Human activity recognition (HAR), which has widely been implemented in various fields which include healthcare, video surveillance and education. We therefore present in this paper a human activity recognition system based on scale and rotation while employing the Kinect depth sensors to obtain the human skeleton joints. In contrast to previous approaches that use joint angles, in this paper we propose that each limb has an angle with the X, Y, Z axes which we employ as feature vectors. The use of the joint angles makes our system scale invariant. We further calculate the body relative direction in the egocentric coordinates in order to provide the rotation invariance. For the system parameters, we employ 8 limbs with their corresponding angles each having the X, Y, Z axes from the coordinate system as feature vectors. The extracted features are finally trained and tested with the Long short term memory (LSTM) Network which gives us an average accuracy of 98.3%.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제2권7호
• /
• pp.479-486
• /
• 2013

정보 시스템의 골격이라고도 할 수 있는 데이터 모델은 프로세스 모델과 함께 정보 시스템에 있어서 중요한 두개의 축을 이룬다. 데이터 모델의 핵심요소로는 엔터티, 속성, 관계가 있으며, 이 중에서도 엔터티는 데이터 모델에서 가장 근본적인 요소로서, 엔터티를 명확하게 도출하지 못한다면 데이터 모델 전체가 모호해지게 된다. 본 연구에서는 엔터티 도출에 대해서만 다루었다. 기존의 엔터티를 도출하는 방법은 설계자의 경험과 업무지식에 많이 의존되고 명확한 절차가 제시되지 않아, 초보자나 미숙련자가 접근하기에는 많은 어려움이 있다. 이를 해결하는데 도움이 될 수 있도록 본 연구에서는 선행연구의 제안 절차를 통해 미리 도출된 대상업무에서 체계적인 절차로 엔터티를 도출할 수 있는 업무기반 엔터티 도출 절차를 제안한다. 그리고 데이터 모델링에 경험이 없는 학부생들을 대상으로 가상업무에 대하여 제안 절차를 적용하도록 하였고, 기존 방법과 제안 절차의 비교가 불가능하다는 점을 감안하여 학생들이 도출한 엔터티와 모범 답안 간의 유사도 검사로 제안 절차를 검증하였다. 그 결과, 모범 답안에 상당히 근접하게 엔터티를 도출한 것을 확인하였다. 따라서 본 연구에서 제안한 절차가 데이터 모델링에 경험이 없는 초보자가 익숙하지 않은 업무에 적용하더라도 모범 답안에 근접하게 엔터티를 도출할 수 있음이 확인되었다. 엔터티를 제외한 데이터 모델의 핵심요소인 속성과 관계의 도출에 대한 연구는 차후로 미룬다.