Real-Time AI-Driven Hazard Detection: Integrating Computer Vision and Sensor Networks for Enhanced Mining Safety

International Journal of Scientific Research in Computer Science Engineering and Information Technology, Journal Year: 2025, Volume and Issue: 11(1), P. 195 - 202

Published: Jan. 3, 2025

This article presents a comprehensive analysis of real-time hazard detection systems in mining operations through the integration computer vision and sensor networks. The explores how artificial intelligence advanced monitoring technologies are transforming traditional safety protocols, introducing innovative solutions for early emergency response. examines implementation sophisticated model architectures video analytics, multilayered networks, data frameworks that enable precise tracking worker behavior, equipment proximity, environmental conditions. Through detailed investigation system performance metrics, challenges, validation processes, this demonstrates significant impact AI-driven on reducing workplace incidents improving operational efficiency. also addresses critical challenges underground environments, including factors, technical constraints, quality management, while providing insights into future developments best practices industry adoption. approach to represents advancement protecting maintaining productive operations.

Language: Английский

---

Automated recognition of construction worker activities using multimodal decision-level fusion

Automation in Construction, Journal Year: 2025, Volume and Issue: 172, P. 106032 - 106032

Published: Feb. 7, 2025

Language: Английский

---

Occlusion-Aware Worker Detection in Masonry Work: Performance Evaluation of YOLOv8 and SAMURAI

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(7), P. 3991 - 3991

Published: April 4, 2025

This study evaluates the performance of You Only Look Once version 8 (YOLOv8) and a SAM-based unified robust zero-shot visual tracker with motion-aware instance-level memory (SAMURAI) for worker detection in masonry construction environments under varying occlusion conditions. Computer vision-based monitoring systems are widely used construction, but traditional object models struggle occlusion, limiting their effectiveness real-world applications. The research employed structured experimental framework to assess both brick transportation laying tasks across three levels: non-occlusion, partial severe occlusion. Results demonstrate that while YOLOv8 processes frames 2.5 3.5 times faster (28–32 FPS versus 9–12 FPS), SAMURAI maintains significantly higher accuracy, particularly conditions (92.67% 52.67%). YOLOv8’s frame-by-frame processing results substantial degradation as severity increases, whereas SAMURAI’s memory-based tracking mechanism enables persistent identification frames. comparative analysis provides valuable insights selecting appropriate technologies based on specific site requirements. is suitable characterized by minimal occlusions high demand real-time detection, more applicable scenarios frequent require sustained activity. selection an model should be initial assessment environmental factors such layout complexity, density, expected frequency. findings contribute advancement reliable enhancing productivity safety management dynamic settings.

Language: Английский

---

Analysis of masonry work activity recognition accuracy using a spatiotemporal graph convolutional network across different camera angles

Automation in Construction, Journal Year: 2025, Volume and Issue: 175, P. 106178 - 106178

Published: April 9, 2025

Language: Английский

---

The Relationship Between Artificial Intelligence (AI) and Building Information Modeling (BIM) Technologies for Sustainable Building in the Context of Smart Cities

Sustainability, Journal Year: 2024, Volume and Issue: 16(24), P. 10848 - 10848

Published: Dec. 11, 2024

The development of information technologies has been exponentially applied to the architecture, engineering, and construction (AEC) industries. extent literature reveals that two most pertinent are building modeling (BIM) artificial intelligence (AI) technologies. radical digitization AEC industry, enabled by BIM AI, contributed emergence “smart cities”, which uses technology improve urban operational sustainable efficiency. Few studies have investigated roles AI in from perspective buildings assisting designers make decisions at city levels. Therefore, purpose this paper is explore research status future trends relationship between BIM-aided context smart provide researchers, designers, developers with potential directions. This adopted a macro micro bibliographic method, used map out general landscape. followed more in-depth analysis fields design, construction, development, life cycle assessment (LCA). results show combination helps optimal on materials, cost, energy, scheduling, monitoring promotes both technical human aspects so achieve Sustainable Development Goals 7 (ensuring access affordable, reliable, modern energy for all), 9 (building resilient infrastructure, promote inclusive industries, foster innovation), 11 inclusive, safe, risk-resilient, cities settlements), 12 consumption production patterns). In addition, BIM, LCA offers great performance, integration should not only consider sustainability but also human-centered design concept health, safety, comfort stakeholders as one goals realize multidimensional based model.

Language: Английский

---

A Voxel-Based 3D reconstruction and action recognition method for construction workers

Advanced Engineering Informatics, Journal Year: 2025, Volume and Issue: 65, P. 103203 - 103203

Published: Feb. 14, 2025

Language: Английский

---

CSOD-24: Construction Site Object Detection Dataset for Safety Monitoring at Construction Site using Deep Learning

Journal of Innovative Image Processing, Journal Year: 2025, Volume and Issue: 7(1), P. 182 - 206

Published: March 1, 2025

Monitoring the use of personal protective equipment (PPE) and worker proximity to heavy machinery are two areas where ensuring safety compliance on construction sites continues be difficult. The lack dynamic ambient circumstances, comprehensive annotations, real-time video data in existing datasets restricts their applicability real-world situations. In order fill these gaps, this work presents CSOD-24, a dataset intended for site object detection monitoring. includes 100 ten-second clips (16.6 minutes total), covering four major classes: "Dump Truck", "Worker with Helmet", without Helmet" "Excavator". videos were recorded at 10 frames per second (fps) annotated .txt, .json, .xml formats. This supports development validation algorithms automated monitoring, detection, tracking environments. CSOD-24 address challenges, enabling robust foundation advancing computer vision-based thereby contributing reduced workplace hazards improved operational efficiency.

Language: Английский

---

Evaluating the effects of safety incentives on worker safety behavior control through image-based activity classification

Frontiers in Public Health, Journal Year: 2024, Volume and Issue: 12

Published: Aug. 12, 2024

Introduction Construction worker safety remains a major concern even as task automation increases. Although incentives have been introduced to encourage compliance, it is still difficult accurately measure the effectiveness of these measures. A simple count accident rates and lower numbers do not necessarily mean that workers are properly complying with regulations. To address this problem, study proposes an image-based approach monitor moment-by-moment behavior evaluate effects different incentive scenarios. Methods By capturing workers’ behaviors using model integrated OpenPose spatiotemporal graph convolutional network, evaluated safety-incentive scenarios on compliance rules while job. The in were designed 1) varying type (i.e., providing rewards penalties) 2) frequency feedback about ones’ own status during tasks. compared average three regulations personal protective equipment self-monitoring hazard avoidance, arranging hook) for each scenario. Results results show rewarding good-compliance more effective when there no status, penalizing non-compliance feedbacks Discussion This provides accurate assessment their by focusing safe promote among construction workers.

Language: Английский

---

Construction Activity Recognition Method Based on Object Detection, Attention Orientation Estimation, and Person Re-Identification

Buildings, Journal Year: 2024, Volume and Issue: 14(6), P. 1644 - 1644

Published: June 3, 2024

Recognition and classification for construction activities help to monitor manage workers. Deep learning computer vision technologies have addressed many limitations of traditional manual methods in complex environments. However, distinguishing different workers establishing a clear recognition logic remain challenging. To address these issues, we propose novel activity method that integrates multiple deep algorithms. complete this research, created three datasets: 727 images entities, 2546 posture orientation estimation, 5455 worker re-identification. First, YOLO v5-based model is trained detection. A person re-identification algorithm then introduced distinguish by tracking their coordinates, body head orientations, postures over time, estimating attention direction. Additionally, object detection developed identify ten common entity objects. The worker’s determined combining attentional orientation, positional information, interaction with detected entities. Ten video clips are selected testing, total 745 instances detected, achieving an accuracy rate 88.5%. With further refinement, shows promise broader application recognition, enhancing site management efficiency.

Language: Английский

---

Achieving On-Site Trustworthy AI Implementation in the Construction Industry: A Framework Across the AI Lifecycle

Buildings, Journal Year: 2024, Volume and Issue: 15(1), P. 21 - 21

Published: Dec. 25, 2024

In recent years, the application of artificial intelligence (AI) technology in construction industry has rapidly emerged, particularly areas such as site monitoring and project management. This demonstrated its great potential enhancing safety productivity construction. However, concerns regarding technical maturity reliability, safety, privacy implications have led to a lack trust AI among stakeholders end users industry, which slows intelligent transformation for on-site implementation. paper reviews frameworks system design across various sectors government regulations requirements achieving trustworthy responsible AI. The principles are then determined. Furthermore, lifecycle framework specifically tailored systems deployed is proposed. addresses six key phases, including planning, data collection, algorithm development, deployment, maintenance, archiving, clarifies development priorities needed each phase enhance trustworthiness acceptance. provides guidance implementation applications, aiming facilitate industry.

Language: Английский

---