Jul 2026· Measurement science and technology· Vol 37, pp. 335102· 0 citations· 37 references
Physics
TL;DR
To ensure the safety of personnel working in coal-fired power generation corridors, a out-of-bounds detection method based on LiDAR and camera fusion was proposed and a out-of-bounds warning and alarm mechanism based on a logical decision tree was designed.
Abstract
Energy and power systems are fundamental to modern society, particularly to the development of the AI industry. Coal-fired power plants operate in complex and hazardous environments that pose substantial risks to personnel safety. To ensure the safety of personnel working in coal-fired power generation corridors, a out-of-bounds detection method based on LiDAR and camera fusion was proposed. To address the issue of sensor attitude offset, IMU data was used to horizontally correct the point cloud. Occlusion detection and clustering algorithms were then combined to extract dynamic bounding boxes for personnel. To accurately identify the spatial relationship between key human body parts and hazardous areas, a 3D joint localization algorithm based on image-point cloud fusion was constructed. Furthermore, a out-of-bounds warning and alarm mechanism based on a logical decision tree was designed. This mechanism analyzes behavioral intent by combining personnel orientation and joint distribution, and uses a ray method to determine whether a joint has entered a hazardous area. Field measurements demonstrated that the system achieved an 80% warning accuracy rate, an 87.3% out-of-bounds alarm accuracy rate, and an average response delay of 89.25 ms; meanwhile, the single-frame error of the 3D joint localization algorithm was controlled within 0.073 m, validating the effectiveness and practicality of this method.
This study proposes an Advanced Surveillance Framework that makes use of YOLOv10, a next-generation real-time object detection algorithm that greatly outperforms conventional single-sensor approaches in precision, recall, and real-time responsiveness.
Sadiya Begum, Lubna Nausheen, Ruqiya Fatima· International Journal of Eng...· 0 citations
High-voltage cables represent a critical trend in future urban development, and the identification of potential hazards in cable channels has become a key focus of operational maintenance research. Traditional manual inspection methods are inefficient and struggle to comprehensively cover complex environments. This research leverages UAV inspection technology and image-based hazard identification algorithms, utilizing a dataset collected from real-world environments. By comparing mainstream object detection algorithms such as Cascade RCNN, Faster RCNN, YOLOv5, YOLOv7, and YOLOv8, the study ultimately adopts YOLOv8 and its segmentation variant—YOLOv8-seg—as the core models. These models achieve high average precision in detecting four major categories of hazards: construction machinery, ancillary facilities, cable terminal platform equipment, and vegetation encroachment. The proposed solution provides an efficient approach for enhancing power system maintenance.
Wei Zhang, Hai Li, Xinyue Liu et al.· Advances in Engineering Tech...· 0 citations
Fire hazards pose a significant threat to human safety and environmental sustainability, particularly in densely populated and infrastructure-critical regions. This paper presents the design and experimental evaluation of a low-cost autonomous fire detection and suppression system based on a multi-sensor fusion approach, aimed at smart and sustainable environments. The proposed system integrates flame, smoke (MQ-2), temperature (DHT22), and ultrasonic sensors to enable reliable fire detection, environmental monitoring, and autonomous navigation. A decision-based sensor fusion algorithm is employed to minimize false alarms and improve detection robustness under varying conditions. The system is capable of autonomously locating fire sources and performing targeted suppression using a servo-controlled water nozzle. Experimental validation conducted across multiple indoor scenarios demonstrates a detection accuracy of 91.2%, an average response time of 2.4 s, and a 30% reduction in false alarms compared to single-sensor methods. The proposed solution offers a cost-effective and scalable approach for early-stage fire response and can be extended to IoT-enabled smart safety systems for sustainable infrastructure.
Deep Singh, Archisman Ghosh, Disha Biswas et al.· 2026 7th International Confe...· 0 citations
With the advancement of the intellectualization of power systems, visible light images captured by inspection robots have found extensive applications in the condition monitoring of transmission lines. However, two key challenges persist in real-world scenarios. First, critical components such as insulators and spacer dampers are frequently configured in dense and small-scale arrangements, rendering them prone to being overlooked by traditional detection methods owing to insufficient feature extraction or inadequate contextual modeling. Second, transmission corridors are often susceptible to various safety hazards, including encroaching vegetation, accumulated water, floating plastic films, and smoke from nearby fires, each posing a significant threat to operational reliability and the safe and stable operation of the power grid. Therefore, we propose a model named Mobile-RCNN that is capable of detecting both densely arranged and small-scale components as well as safety hazards. First, we constructed a safety hazard dataset through on-site photography. Second, we introduced the Q-linear-NMS post-processing method that replaced the non-maximum suppression (NMS) approach with Soft-NMS and incorporated a linear suppression control coefficient. Then, we replaced the intersection over union (IOU) evaluation method for overlapping detection boxes with the distance-IOU (DIOU) evaluation method and introduced an adjustment parameter for the degree of score suppression. Finally, to enable the backbone feature extraction network to better learn the characteristics of safety hazards and provide rich information for the subsequent detection network, we adopted MobileNetV3 as the feature extraction network for the Mobile-RCNN. The experimental results demonstrated that the proposed model detected densely arranged targets and accurately identified the locations of safety hazards, providing robust technical support for the construction of an intelligent, reliable, and all-weather operation and maintenance system for power systems.