Aug 2026· International Journal of Engineering Research and Science & Technology· 0 citations
TL;DR
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.
Abstract
Deep learning-based object identification models have been incorporated for improved situational awareness and security monitoring as a result of the quick development of intelligent surveillance systems. In order to detect dangerous objects including guns, explosives, and suspicious goods in complicated situations, this study proposes an Advanced Surveillance Framework that makes use of YOLOv10, a next-generation real-time object detection algorithm. In order to increase detection accuracy in a variety of illumination and occlusion scenarios, the suggested system integrates visual and infrared modalities through multi-sensor data fusion. Through enhanced feature aggregation, adaptive anchor mechanisms, and transformer-based attention modules, YOLOv10's optimised architecture provides greater speed–accuracy trade-offs. The fusion-based YOLOv10 model is a potent solution for contemporary surveillance applications in public safety, border control, and smart city security networks because experimental results show that it greatly outperforms conventional single-sensor approaches in precision, recall, and real-time responsiveness.
Weapon-related threats in public places demand intelligent surveillance systems capable of detecting dangerous objects accurately and in real time. This paper presents DeepGuard, an intelligent deep learning framework for automated weapon detection in images and surveillance videos using Faster Region-Based Convolutional Neural Network (Faster R-CNN) and Single Shot Detector (SSD). The proposed framework employs annotated weapon datasets for model training and utilizes convolutional neural networks to identify and localize weapons with bounding boxes. A comparative evaluation of SSD and Faster RCNN is conducted to analyze their detection accuracy and inference speed. Experimental results demonstrate that Faster R-CNN achieves superior detection accuracy, whereas SSD provides faster processing suitable for real-time applications. The developed system effectively identifies weapons in diverse surveillance environments, enhancing public safety through early threat detection and continuous monitoring. The proposed framework offers a reliable, scalable, and intelligent solution for smart surveillance systems, making it suitable for deployment in airports, railway stations, educational institutions, commercial buildings, and other high-security environments.
Chengoli prashanth, Sk.Mahammadunnisa· American Journal of AI Cyber...· 0 citations
Rapid urbanization has increased the need for surveillance systems that can monitor multiple public safety risks at the same time. Traditional systems often use separate solutions for facial recognition, vehicle identification, fire detection, and behavioral analysis, resulting in fragmented infrastructure and multiple interfaces for operators to manage. This paper presents City Sentinel, a unified AI-based surveillance framework that integrates six detection capabilities into one scalable platform: facial recognition, automatic number plate recognition (ANPR), fire and smoke detection, weapon and knife detection, violence detection, and road accident detection. The system combines a Next.js operator dashboard, FastAPI backend, cloud-based PostgreSQL event storage, InsightFace and YOLOv8 vision models, and EasyOCR for plate recognition. Camera streams are processed through dedicated inference workers using RTSP. On a workstation equipped with an NVIDIA RTX 3060 GPU, the system achieves a median end-to-end latency of 743 ms and supports four concurrent RTSP streams within a two-second latency limit. It achieves a 91.2% face-match rate, 85.7% plate-reading accuracy, and mAP@0.5 scores of 0.846 to 0.889 across the fire, knife, and weapon detection modules. In user-acceptance testing, operators could enroll a new identity in under one minute and identify a flagged person from live footage in an average of 12 seconds. The results demonstrate that a modular, open-source, multi-model architecture can provide broad surveillance coverage, cloud-based auditability, and flexibility for adding new detection capabilities while maintaining practical real-time performance.
Hanan Syed Shabir, Noor Fatima, Safia Baloch et al.· 0 citations
Real-time weapon detection in video surveillance systems is a critical requirement for proactive security applications, particularly under the computational and latency constraints imposed by edge artificial intelligence deployments. While the YOLO family of object detectors has undergone continuous architectural evolution, the recently introduced YOLOv26 represents a significant redesign aimed at improving efficiency, stability, and deployment suitability across a wide range of hardware platforms. This work presents a comprehensive and homogeneous experimental evaluation of the full YOLOv26 model family, ranging from nano (YOLOv26n) to extra-large (YOLOv26x) variants, for real-time weapon detection in surveillance imagery. All models are trained and evaluated under identical conditions using a dataset that explicitly includes visually similar non-weapon objects as hard negatives, enabling a realistic assessment of false positives and false negatives in safety critical scenarios. The analysis encompasses training and validation dynamics, precision, recall evolution, mean Average Precision (mAP) at multiple IoU thresholds, class-wise confusion matrices, and inference latency. Results show that performance improves consistently from smaller to medium sized models, with YOLOv26m achieving the most balanced trade off between detection accuracy, robustness, and computational cost. Larger variants provide marginal accuracy gains at significantly higher complexity, revealing diminishing returns for edge oriented deployments. Overall, the findings demonstrate that the YOLOv26 architecture offers a scalable and mature detection framework, where model selection can be guided by explicit operational criteria rather than raw accuracy alone. This study establishes a strong baseline for future work on real world edge deployment, multi camera surveillance systems, and hardware aware optimization of next generation YOLO detectors.
Carlos Silva, Carolina Del-Valle-Soto, J. Varela-Aldás· 2026 6th International Confe...· 0 citations
Intelligent surveillance systems need very dynamic visual intelligence that will recognise individuals and mark the presence of possible threats in real-time video surveillance. To overcome this requirement, a single deep learning system called SmartVision-AI is proposed, which combines face recognition and weapon identification in one feature-based architecture. The approach uses a dual-branch convolutional encoder, attention-based feature fusion, and multi-task learning, which is optimized towards low-latency CCTV systems. Tests on mixed-face and weapon image data sets show that the architecture can deliver face recognition accuracy of 96.8%, multi-class weapon detection accuracy of 94.7%, a false-positive rate drop to 21%, a precision increase of up to 19%, and a processing time of only 38 ms/frame, which can be effectively deployed in near real-time. Further processing indicates that there is an increase in temporal stability by 32% and a reduction in the use of GPU memory by 27% in comparison with individual task-specific models. The findings attest to the fact that SmartVision-AI provides a powerful, effective, and scalable intelligent threat-aware video surveillance system.
P.Shobana, V.Sai Shanmuga Raja, P. S. Rajakumar et al.· International journal of com...· 0 citations
The design of a low-cost driver assistance system (DAS) using monocular camera input and artificial intelligence to enhance road awareness consists of using low-cost sensors instead of costly configurable sensors used in typical systems. The hybrid perception architecture of this system incorporates deep learning (via optimized YOLOv8) and traditional computer vision techniques to achieve high accuracy in detecting vehicles and pedestrians, which is consistent regardless of traffic conditions. Additionally, the hybrid lane detection algorithm combines edge-filtering techniques with geometric models to allow for lane detection in low-light or poorly marked lane conditions. Also, the development of a modular processing pipeline allows for real-time video preprocessing, feature extraction and risk assessment, therefore requiring less computational resources than standard DAS systems. Finally, testing showed that this DAS system consistently performs in real-time and achieves an acceptable degree of accuracy, irrespective of environmental conditions. The DAS system provides a common structure for a variety of vision techniques and can be scaled and constructed for a lower cost than most current DAS solutions, thereby facilitating the development of intelligent transportation systems and increasing access to transportation technology.
Poola Joshika, C. Dharshana, Shreya Sridharan et al.· 2026 6th International Confe...· 0 citations
A proactive, real-time computer vision system designed to detect potentially suspicious behavior around parked vehicles, with a specific focus on unauthorized proximity and loitering is proposed, making it a strong candidate for practical urban vehicle monitoring, subject to further large-scale validation across diverse environments.
Umar Adeel, Ammar Rashid, S. Yusof et al.· Information· 0 citations