• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권12호
• /
• 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.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권3호
• /
• pp.701-720
• /
• 2023

Aerial objects are more challenging to segment than normal objects, which are usually smaller and have less textural detail. In the process of segmentation, target objects are easily omitted and misdetected, which is problematic. To alleviate these issues, we propose local aggregation feature pyramid networks (LAFPNs) and pyramid integrated context modules (PICMs) for aerial object segmentation. First, using an LAFPN, while strengthening the deep features, the extent to which low-level features interfere with high-level features is reduced, and numerous dense and small aerial targets are prevented from being mistakenly detected as a whole. Second, the PICM uses global information to guide local features, which enhances the network's comprehensive understanding of an entire image and reduces the missed detection of small aerial objects due to insufficient texture information. We evaluate our network with the MS COCO dataset using three categories: airplanes, birds, and kites. Compared with Mask R-CNN, our network achieves performance improvements of 1.7%, 4.9%, and 7.7% in terms of the AP metrics for the three categories. Without pretraining or any postprocessing, the segmentation performance of our network for aerial objects is superior to that of several recent methods based on classic algorithms.

• 융합보안논문지
• /
• 제22권5호
• /
• pp.3-9
• /
• 2022

최근 들어 4차산업 연관기술인 사이버물리시스템(CPS) 구축을 위해 물리 모델과 제어회로 시뮬레이션을 위한 가상 제어시스템 구축 작업이 다양한 산업 분야에서 요구가 점점 증가하고 있다. 전자 문서화 되지 않은 문서들에 대한 직접입력을 통한 변환은 시간과 비용이 많이 소모된다. 이를 위해 이미 출력된 대량의 도면을 인공지능을 이용한 객체 인식을 통해 디지털화 작업은 매우 중요하다고 할 수 있다. 본 논문에서는 도면내 객체를 정확하게 인식하고 이를 다양한 응용에 활용할 수 있도록 하기 위하여 도면내 객체의 특징을 분석하여 인공지능을 활용한 인식 기법을 제안하였다. 객체 인식의 성능을 높이기 위하여 객체별 인식 후 그 정보를 저장하는 중간 파일을 생성하게 하였다. 그리고 인식 결과를 도면에서 삭제하여 다음 인식 대상의 인식률을 향상시켰다. 그리고 그 인식 결과를 표준화 포맷 문서로 저장하여 이를 제어시스템의 다양한 분야에 활용할 수 있도록 하였다. 본 논문에서 제안한 기법의 우수한 성능은 위해 실험을 통해 확인할 수 있었다.

• 국제학술발표논문집
• /
• The 9th International Conference on Construction Engineering and Project Management
• /
• pp.877-885
• /
• 2022

Vision-based activity recognition has been widely attempted at construction sites to estimate productivity and enhance workers' health and safety. Previous studies have focused on extracting an individual worker's postural information from sequential image frames for activity recognition. However, various trades of workers perform different tasks with similar postural patterns, which degrades the performance of activity recognition based on postural information. To this end, this research exploited a concept of human-object interaction, the interaction between a worker and their surrounding objects, considering the fact that trade workers interact with a specific object (e.g., working tools or construction materials) relevant to their trades. This research developed an approach to understand the context from sequential image frames based on four features: posture, object, spatial features, and temporal feature. Both posture and object features were used to analyze the interaction between the worker and the target object, and the other two features were used to detect movements from the entire region of image frames in both temporal and spatial domains. The developed approach used convolutional neural networks (CNN) for feature extractors and activity classifiers and long short-term memory (LSTM) was also used as an activity classifier. The developed approach provided an average accuracy of 85.96% for classifying 12 target construction tasks performed by two trades of workers, which was higher than two benchmark models. This experimental result indicated that integrating a concept of the human-object interaction offers great benefits in activity recognition when various trade workers coexist in a scene.

• Advances in concrete construction
• /
• 제14권3호
• /
• pp.185-194
• /
• 2022

For producing cement and concrete, the construction field has been encouraged by the usage of industrial soil waste (or) secondary materials since it decreases the utilization of natural resources. Simultaneously, for ensuring the quality, the analyses of the strength along with durability properties of that sort of cement and concrete are required. The prediction of strength along with other properties of High-Performance Concrete (HPC) by optimization and machine learning algorithms are focused by already available research methods. However, an error and accuracy issue are possessed. Therefore, the Enhanced Deep Neural Network (EDNN) based strength along with durability prediction of HPC was utilized by this research method. Initially, the data is gathered in the proposed work. Then, the data's pre-processing is done by the elimination of missing data along with normalization. Next, from the pre-processed data, the features are extracted. Hence, the data input to the EDNN algorithm which predicts the strength along with durability properties of the specific mixing input designs. Using the Switched Multi-Objective Jellyfish Optimization (SMOJO) algorithm, the weight value is initialized in the EDNN. The Gaussian radial function is utilized as the activation function. The proposed EDNN's performance is examined with the already available algorithms in the experimental analysis. Based on the RMSE, MAE, MAPE, and R² metrics, the performance of the proposed EDNN is compared to the existing DNN, CNN, ANN, and SVM methods. Further, according to the metrices, the proposed EDNN performs better. Moreover, the effectiveness of proposed EDNN is examined based on the accuracy, precision, recall, and F-Measure metrics. With the already-existing algorithms i.e., JO, GWO, PSO, and GA, the fitness for the proposed SMOJO algorithm is also examined. The proposed SMOJO algorithm achieves a higher fitness value than the already available algorithm.

• 국제학술발표논문집
• /
• The 8th International Conference on Construction Engineering and Project Management
• /
• pp.87-95
• /
• 2020

Generally, occupational safety and particularly construction safety is an intricate phenomenon. Industry professionals have devoted vital attention to enforcing Occupational Safety and Health (OHS) from the last three decades to enhance safety management in construction. Despite the efforts of the safety professionals and government agencies, current safety management still relies on manual inspections which are infrequent, time-consuming and prone to error. Extensive research has been carried out to deal with high fatality rates confronting by the construction industry. Sensor systems, visualization-based technologies, and tracking techniques have been deployed by researchers in the last decade. Recently in the construction industry, computer vision has attracted significant attention worldwide. However, the literature revealed the narrow scope of the computer vision technology for safety management, hence, broad scope research for safety monitoring is desired to attain a complete automatic job site monitoring. With this regard, the development of a broader scope computer vision-based risk recognition system for correlation detection between the construction entities is inevitable. For this purpose, a detailed analysis has been conducted and related rules which depict the correlations (positive and negative) between the construction entities were extracted. Deep learning supported Mask R-CNN algorithm is applied to train the model. As proof of concept, a prototype is developed based on real scenarios. The proposed approach is expected to enhance the effectiveness of safety inspection and reduce the encountered burden on safety managers. It is anticipated that this approach may enable a reduction in injuries and fatalities by implementing the exact relevant safety rules and will contribute to enhance the overall safety management and monitoring performance.

• 한국항행학회논문지
• /
• 제27권6호
• /
• pp.776-781
• /
• 2023

K-UAM은 2035년까지의 성숙기 이후 상용화될 예정이다. UAM 회랑은 기존의 헬리콥터 회랑을 수직 분리하여 사용될 예정이기에 회량 사용량이 증가할 것으로 예상된다. 따라서 회랑을 모니터링하는 시스템도 필요하다. 최근 객체 검출 알고리즘이 크게 발전하였다. 객체 검출 알고리즘은 1단계 탐지와, 2단계 탐지 모델로 나뉜다. 실시간 객체 검출에 있어서 2단계 모델은 너무 느리기에 적합하지 않다. 기존 1단계 모델은 정확도에 문제가 있었지만, 버전 업그레이드를 통해 성능이 향상되었다. 1단계 모델 중 YOLO-V5는 모자이크 기법을 통한 소형 객체 검출 성능을 향상시킨 모델이다. 따라서 YOLO-V5는 넓은 회랑의 실시간 모니터링에 가장 적합하다고 판단된다. 본 논문에서는 YOLO-V5 알고리즘을 학습시켜 궁극적으로 회랑 모니터링 시스템에 대한 적합도를 분석한다.

• 한국전자통신학회논문지
• /
• 제18권6호
• /
• pp.1321-1330
• /
• 2023

본 연구는 증강현실에서 적용할 캐릭터 생성에서 단일 이미지를 통해 여러 객체에 대한 3D 자세 추정 문제를 연구한다. 기존 top-down 방식에서는 이미지 내의 모든 객체를 먼저 감지하고, 그 후에 각각의 객체를 독립적으로 재구성한다. 문제는 이렇게 재구성된 객체들 사이의 중첩이나 깊이 순서가 불일치 하는 일관성 없는 결과가 발생할 수 있다. 본 연구의 목적은 이러한 문제점을 해결하고, 장면 내의 모든 객체에 대한 일관된 3D 재구성을 제공하는 단일 네트워크를 개발하는 것이다. SMPL 매개변수체를 기반으로 한 인체 모델을 top-down 프레임워크에 통합이 중요한 선택이 되었으며, 이를 통해 거리 필드 기반의 충돌 손실과 깊이 순서를 고려하는 손실 두 가지를 도입하였다. 첫 번째 손실은 재구성된 사람들 사이의 중첩을 방지하며, 두 번째 손실은 가림막 추론과 주석이 달린 인스턴스 분할을 일관되게 렌더링하기 위해 객체들의 깊이 순서를 조정한다. 이러한 방법은 네트워크에 이미지의 명시적인 3D 주석 없이도 깊이 정보를 제공하게 한다. 실험 결과, 기존의 Interpenetration loss 방법은 MuPoTS-3D가 114, PoseTrack이 654에 비해서 본 연구의 방법론인 Lp 손실로 네트워크를 훈련시킬 때 MuPoTS-3D가 34, PoseTrack이 202로 충돌수가 크게 감소하는 것으로 나타났다. 본 연구 방법은 표준 3D 자세벤치마크에서 기존 방법보다 더 나은 성능을 보여주었고, 제안된 손실들은 자연 이미지에서 더욱 일관된 재구성을 실현하게 하였다.