Search Results (637)

This paper proposes a post-training structured channel pruning scheme leveraging Gradient-Weighted Class Activation Mapping (Grad-CAM) for few-class fire detection under limited computational resources. After standard training, category-specific gradient signals extracted from the detection heads are used to generate channel-wise class activation maps over multi-scale feature layers, and fire and smoke responses are fused to rank channel importance. A layer-level retention quota is further applied to implement structured channel pruning, followed by lightweight fine-tuning. The pipeline does not require additional sparsity-inducing training. We validate the method on a self-established fire and smoke dataset containing 9041 images, using YOLOv5s, YOLOv5m, and YOLOv5l as baseline detectors. In workflow-level comparisons, the proposed method achieved higher mAP@0.5 than the implemented L1-based workflow at 40% and 60% pruning, but not at 80%. At 60% pruning, the pruned YOLOv5s model contained 2.158 M parameters and required 2.901 GFLOPs. These results indicate that Grad-CAM provides a useful class-aware criterion for channel importance and offers a promising model-compression strategy for resource-constrained fire detection, although physical edge-device performance remains to be evaluated. Full article

Wildfires represent a major environmental hazard with significant impacts on ecosystems, climate, biodiversity, and human activities. The increasing frequency and intensity of wildfire events have highlighted the need for reliable and timely detection techniques based on satellite remote sensing. This study investigates the application of Proper Orthogonal Decomposition (POD) to thermal observations acquired from the Spinning Enhanced Visible and Infrared Imager (SEVIRI) onboard the Meteosat Second Generation (MSG) satellite for wildfire anomaly detection. A wildfire event that occurred on 8 August 2021 in Calabria, Southern Italy, was selected as the primary case study. To assess the consistency of the POD response beyond the primary case, the analysis was further extended to two additional wildfire events, Viggianello–Abate and Pazzano–Montestella, using the 15 × 15 pixel extraction window. Middle Infrared (MIR, 3.9 $\mathsf{\mu }$m) observations collected at 15 min intervals over a complete day were analyzed using four different spatial extraction windows (3 × 3, 15 × 15, 30 × 30, and 45 × 45 pixels). POD was employed to separate dominant background thermal variability from localized fire-induced anomalies. The analysis focused on higher-order POD modes, particularly the 6th, 7th, and 8th modes, which exhibited enhanced sensitivity to wildfire activity. Results showed that POD successfully identified thermal anomalies corresponding to wildfire occurrence times independently detected by the RST-FIRES methodology. The comparison of extraction window sizes revealed that the 15 × 15 pixel window provided the best balance between anomaly enhancement, spatial localization, and noise reduction. Larger windows introduced excessive spatial smoothing and reduced localization capability, whereas the smallest window was more affected by noise. The findings demonstrate the potential of POD as an effective complementary approach for wildfire detection and monitoring using geostationary satellite observations. Full article

Tiny fire spot detection in UAV images under complex forest backgrounds remains challenging due to tiny target size, sparse distribution, weak feature responses, and background interference. This paper proposes a Multi-Scale Feature Enhancement Network (MSF-YOLO) for tiny fire spot detection. Specifically, a lightweight C2f-ARG module is designed by integrating Ghost feature generation and channel recalibration mechanisms to enhance weak fire spot representation while reducing redundant features. A C2f-LGPA module is designed to model local fine-grained information and global contextual dependencies, improving target discrimination under complex forest environments. Additionally, a P2 tiny-object detection branch is incorporated to preserve spatial details and enhance the perception capability of tiny targets. A UAV forest fire spot detection dataset was constructed, and extensive experiments were conducted. Experimental results demonstrate that MSF-YOLO achieves a Recall of 79.27, representing an improvement of 5.09% over the baseline YOLOv8s. The mAP@0.5 and mAP@0.5:0.95 values are improved by 3.99% and 5.35%, respectively. Moreover, compared with eight improved YOLO-based small-object detectors, MSF-YOLO achieves superior overall detection performance, with a 1.35% improvement in mAP@0.5 over the best-performing comparison method. The proposed MSF-YOLO effectively addresses the challenge of early-stage tiny fire spot detection in forest fire. Full article

Wastewater treatment plants (WWTPs) are essential infrastructure for urban water management. Their stable operation is critical to effluent quality and public safety. However, wastewater treatment involves complex biochemical processes and extensive electromechanical equipment. Hazards such as sludge flotation in secondary clarifiers, fire, electric shock, and toxic gas poisoning may occur. These hazards can threaten worker safety and reduce treatment efficiency. Traditional inspection mainly relies on manual patrols, fixed-camera monitoring, and experience-based judgment. These methods often have low efficiency, limited coverage, and delayed responses. To address these limitations, this paper investigates robot-based hazard detection for WWTPs. A multisource hazard detection dataset is constructed for secondary clarifiers and confined spaces, including images collected by an inspection robot. Object detection models are then applied to identify typical hazards. Comparative experiments are conducted using Faster R-CNN and several YOLO-series models. YOLOv12 achieves mAP@0.5 values of 0.917 and 0.819 for sludge flotation detection and confined space hazard detection, respectively. It also provides a good balance between detection performance and inference efficiency. The results demonstrate that robot vision combined with object detection can support intelligent inspection in WWTPs. Full article

The Greater Cape Floristic Region (GCFR) in South Africa is a fire-prone biodiversity hotspot where high species richness, structural complexity, and small plant sizes (0.0001–4 m²) pose substantial challenges for remote sensing-based biodiversity assessment. Spectral similarity among species and the mismatch between plant size and sensor pixel dimensions limit the capacity of current and forthcoming spaceborne systems to resolve individual species and accurately detect plot-level diversity changes. We therefore developed a physics-based simulation framework that couples fynbos trait measurements with radiative transfer modeling in the DIRSIG (Digital Imaging and Remote Sensing Image Generation) environment towards quantifying information loss across spectral and spatial scales and to define theoretical limits for biodiversity monitoring. We constructed a three-dimensional virtual scene of post-fire fynbos communities in Grootbos Private Nature Reserve, integrating high-resolution imagery, terrestrial laser scanning (TLS), and structure-from-motion (SfM)-derived point clouds. Field measurements of mean diameter and percent cover were used to scale vegetation models and constrain species abundance. We distributed plant instances using a blue noise sampling algorithm, guided by density maps derived from unmanned aerial system (UAS) imagery. Species-specific optical properties were parameterized using field-measured reflectance data and the PROSPECT radiative transfer model, while terrain structure was derived from SfM-based digital terrain models. The integrated scene was used to simulate multispectral (DJI Mavic 3 MSI), hyperspectral (AVIRIS-NG), and light detection and ranging (LiDAR) observations. Agreement between simulated outputs were evaluated against corresponding field-acquired datasets using spectral signatures and vegetation indices. This framework enables systematic assessment of sensor specification effects on spectral biodiversity metrics and provides a pathway for evaluating theoretical limits of species discrimination across airborne and satellite platforms. Full article

Wildfires need a fast response, and a small unmanned aerial vehicle (UAV) carrying its own detector is an appealing way to find them early. Almost every published UAV fire detector is validated on a single corpus, which leaves open the question that matters to an operator: how much of the reported accuracy survives a change in scene. We trained unmodified YOLOv8n, YOLOv8s, and YOLOv8m on two public benchmarks that ship official test splits, D-Fire and the UAV subset of the Flame and Smoke Detection Dataset, under one fixed recipe with three seeds on D-Fire, then evaluated every checkpoint on both test splits with frozen weights. In domain, YOLOv8m reached 79.04 ± 0.21% mAP@0.5 on D-Fire, within 0.04 points of the published value for the same architecture on the same split, and 92.71% on FASDD_UAV. Moved across corpora, the same weights fell 38 and 65 points below a locally trained model, with fire degrading about twice as far as smoke. A perceptual-hash audit then found near-duplicates of training images in 46.0% and 90.7% of the two test sets; re-evaluating on the uncontaminated remainder costs 2.8 and 12.4 points, closes two-thirds of the apparent difficulty gap between the corpora, and leaves corrected transfer shortfalls near 26 and 62 points. Onboard, an F450 carrying a Pixhawk 2.4.8 and a Jetson Nano A02 flew and detected a controlled fire, but full-resolution inference ran 4.0 to 41.4 times slower than the video it consumed. Both gaps must close before a platform of this class can support operational wildfire monitoring. Full article

As interest in autonomous ships continues to grow, the need for systems capable of automatically detecting and suppressing fires without human intervention is increasing. In this study, a four-channel + UV fire detection model integrating RGB, infrared (IR), and ultraviolet (UV) sensors, along with a LiDAR-based fire source localization system, is proposed. The fire detection model consists of an Xception-based network receiving four-channel (RGB + UV) images and a dense-based network processing UV time series data as inputs. The proposed model achieved high accuracy on the validation dataset, while the inclusion of UV data effectively eliminated false alarms caused by fire-like sources such as welding. For fire source localization, a four-channel YOLOv7 model combined with LiDAR projection was implemented, achieving a localization accuracy that significantly outperformed the conventional triangulation method. The effectiveness of the proposed system was validated through integrated experiments conducted under both open-space and occluded-fire scenarios. Full article

The growing frequency of unauthorized UAV activities has increased the demand for real-time perception and rapid response on resource-constrained edge devices. This study proposes an edge-deployable UAV detection and net-capture system based on Net-Capture Detection YOLO (NCDet-YOLO). Developed from YOLOv8n, NCDet-YOLO incorporates C2f_Faster, SPD_Conv, EMA, and a lightweight three-scale detection head, with CrossKD used to compensate for accuracy loss caused by structural compression. The dataset contains 6615 images and was divided into 5292 training and 1323 validation images. The self-collected data include DJI Phantom 4 and DJI Inspire 2 UAVs observed at approximately 4–30 m under different daytime backgrounds. NCDet-YOLO achieves an mAP50–95 of 0.6504 with 1.55 M parameters and 4.1 GFLOPs. On a Jetson Orin NX Super under the 15 W power mode, it achieves 31.53 FPS, representing a 31.67% increase over YOLOv8n. The detector is further integrated with target alignment, distance determination, trigger control, and net-capture execution. In 10 real-platform trials, 8 captures were successful, corresponding to an 80.0% success rate, with one false-trigger event and an end-to-end latency from target detection to net-capture firing of approximately 400 ms. Full article

Identifying coal fire risk areas is essential for safe production in coal mines. Land Surface Temperature (LST) retrieval and high-temperature anomaly extraction are core techniques for coal fire risk detection. To address the insufficient evaluation of the accuracy of relevant algorithms for arid open-pit mines, this study takes the Xingsheng Open-Pit Coal Mine in Yiwu County, Xinjiang as the research object. Based on Landsat imagery and UAV thermal infrared data, we systematically compared five mainstream LST retrieval algorithms and six high-temperature anomaly extraction algorithms and determined the optimal combination for long-term monitoring. The results indicate that all five algorithms can effectively depict LST spatial distribution under normal temperature conditions. The Jiménez-Muñoz split-window algorithm performs best for small-scale coal fire identification, with a mean absolute error of 3.25 °C and a relative error of 5.53%, and its fitting slope of 0.82 proves superior stability. For high-temperature anomaly extraction methods, the gradient threshold method achieves a 100% overlap rate with actual anomalies and no omission, which is ideal for large-scale surveys; the cluster analysis method balances detection accuracy and economic benefits for pit-scale investigations. Using 52 valid Landsat images from 2013 to 2025, long-term monitoring reveals that high-temperature anomalies are most active in summer, with an average patch area of 5.65 × 10⁵ m², and weaken sharply in winter. According to the observed spatiotemporal evolution patterns, the dynamic changes in thermal anomalies are inferred to be mainly associated with human mining activities, with coal seam conditions as the secondary influencing factor. This study provides reliable technical references for coal mine safety management and coal fire disaster prevention. Full article

Timely wildfire monitoring depends critically on optical and thermal infrared sensor observations from spaceborne instruments. The TS-SatFire benchmark (2025) consolidates multispectral VIIRS image stacks from Suomi-NPP and NOAA-20 for three tasks: active fire (AF) detection, burned area (BA) mapping, and fire progression (FP) prediction. We make two contributions. First, a systematic label-quality audit reveals that many fires lack ground-truth annotations; 18 training fires and 2 test fires were excluded for AF, and the two unannotated test fires cannot be scored by any model. We further document the benchmark’s scoring procedure, which differs from ours in ways that make the two sets of figures incomparable, and the BA label encoding in the released GeoTIFFs; the BA task is only audited. Second, we propose SpaSE-UNet3D, a spatial squeeze-and-excitation 3D U-Net whose spatial-only $(1,3,3)$ convolutions avoid temporal mixing on short observation windows, while SE channel attention reweights the VIIRS spectral bands dynamically. With micro-averaging over all test pixels, it reaches F1 = $0.8549\pm 0.0005$ on AF and $0.3845\pm 0.0221$ on FP at TS = 2, matching or exceeding the strongest published baselines on their respective terms. A single-day AF input reaches $0.8520\pm 0.0008$, within 0.003 of the two-day figure, indicating that one acquisition carries most of the detectable signal, whereas published baselines use up to six days; on FP, we use one third of their temporal context. An ablation shows the spatial-only design matches the accuracy of a full $(3,3,3)$ network with $2.72\times$ fewer parameters. Code and results are publicly available. Full article

Wildfires are a major environmental hazard with severe consequences for ecosystems, air quality, infrastructure, and public safety. The rising incidence and severity of wildfire events worldwide have increased the need for reliable early detection and monitoring systems. Remote sensing technologies, such as satellite and unmanned aerial vehicle (UAV) imagery, along with ground-based Closed-Circuit Television (CCTV) cameras, provide valuable geospatial data for large-scale wildfire monitoring. Recent advances in deep learning, particularly Convolutional Neural Networks (CNNs) and Transformer-based architectures, have significantly improved the accuracy of wildfire detection systems. Despite these advances, balancing local feature representation with global contextual modeling remains challenging. CNNs effectively capture local spatial features but have limited receptive fields, whereas Vision Transformers (ViTs) model long-range dependencies but often overlook fine-grained local details and require substantial computational resources. Consequently, accurately detecting small, occluded, and visually ambiguous fire regions remains difficult, particularly for real-time deployment on resource-constrained edge devices. To address these challenges, this study proposes EdgeNeXt-Attn, an enhanced EdgeNeXt-based framework that effectively integrates local feature learning and global contextual modeling through channel and spatial attention mechanisms. The proposed model improves the detection of small, occluded, and visually ambiguous fire regions while maintaining the computational efficiency required for real-time edge deployment. The proposed framework is evaluated on four multi-platform benchmarks spanning ground-based CCTV (DFAN, Complex-Fire), aerial drone (FLAME), and mixed drone–satellite (ADSF) imagery, achieving 92.09%, 95.16%, 96.65%, and 87.81% accuracy, respectively, and outperforming recent state-of-the-art baselines. With only 5.3M parameters, the model achieves real-time inference at 85.9, 27.3, and 8.4 FPS on GPU, CPU, and Raspberry Pi, respectively. Furthermore, ablation studies and Grad-CAM analysis validate its effectiveness and accurate fire localization. These results demonstrate an accurate and computationally efficient framework for real-time wildfire monitoring using multi-platform remote sensing and ground-based imaging systems. Full article

This paper presents NexFire Pro, a deployable desktop framework for real-time fire and smoke surveillance across multiple simultaneous IP or USB camera streams. The research contribution is a failure-mode-driven pipeline design methodology: each of the five post-detection stages is derived from a distinct, identifiable failure-mode category (illumination drift, static false-positive objects, spatially expected heat sources, temporal noise), evaluated with a leakage-free, source-disjoint held-out protocol that avoids the temporal data leakage that frame-level splits introduce. A fine-tuned YOLO11s backbone serves as the fixed detection substrate, surrounded by a five-stage processing pipeline: (1) CLAHE and adaptive gamma correction; (2) neural detection; (3) MOG2-based motion validation; (4) normalised-coordinate ROI exclusion; and (5) temporal alarm hysteresis. On a leakage-free held-out image test set, the detector attains mAP@0.5 = 0.65 (F1 = 0.70). A leave-one-out ablation on non-fire video shows that motion validation is the dominant false-alarm suppressor, reducing a raw-detector baseline of eight false alarms to zero. The proposed design-and-evaluation framework is intended to be transferable to other reliability-oriented computer vision applications beyond fire detection. Full article

To address the problems of small fire-spot scale, blurred boundaries, complex backgrounds, and insufficient feature representation of weak targets in Unmanned Aerial Vehicle (UAV)-based early forest fire detection, a YOLOv8n-based forest fire detection model, termed TriRHC-YOLO, is proposed. The model first introduces Reparameterized VGG (RepVGG)Block into the backbone network to enhance the extraction capability of shallow local features. Subsequently, a Hierarchical Feature Attention (HFA) module is designed to collaboratively model fire-spot features from three levels, namely directional structures, local textures, and global semantics, thereby enhancing the network’s capability to discriminate fire targets and suppressing interference from complex forest backgrounds. Finally, a Cross Stage Partial Feature Fusion with Cascade Star Block (C2f-CStar) module is designed to improve the representation capability of the model for local structural information and weak salient fire-spot features under complex backgrounds through cascaded spatial feature reconstruction and a star-shaped multiplicative gating mechanism. In addition, a UAV-specific early forest fire detection dataset is constructed based on the FLAME and FLAME_VISION datasets, and experimental validation is conducted on this dataset. The experimental results show that the proposed TriRHC-YOLO outperforms several classical YOLO algorithms, including YOLO11n, YOLO12, and YOLO26, as well as six advanced YOLO-based improved models. The Recall, mean Average Precision (mAP)@0.5, and mAP@0.5:0.95 reach 0.769, 0.848, and 0.608, respectively. The results of the ablation experiments further verify the effectiveness of the three designed modules. Moreover, the proposed model contains only 3.181 M parameters and achieves 168.251 Frames Per Second (FPS), demonstrating favorable real-time detection capability. Overall, the proposed method can effectively improve the detection accuracy of early weak fire targets and the background suppression capability under complex forest backgrounds, making it suitable for real-time UAV-based forest fire inspection tasks. Full article

Forest fires are a major environmental concern in Northern Thailand, contributing to ecosystem degradation, biodiversity loss, and seasonal air pollution. This study identified the environmental factors influencing forest fire occurrence and developed a forest fire susceptibility map using an integrated Geographic Information System (GIS) and Logistic Regression (LR) framework in Chom Thong District, Chiang Mai Province, Thailand. Fire occurrence data were derived from Visible Infrared Imaging Radiometer Suite (VIIRS) active fire hotspots detected by the Suomi National Polar-orbiting Partnership satellite (Suomi-NPP satellite) during 2023–2025. A total of 1674 hotspots were identified (616 in 2023, 889 in 2024, and 169 in 2025). Ten environmental variables, including elevation, slope, aspect, Topographic Wetness Index (TWI), stream density, rainfall, Normalized Difference Vegetation Index (NDVI), Modified Normalized Difference Water Index (MNDWI), Land Surface Temperature (LST), and land-use, were analyzed. The LR model was trained using 2293 training samples (70%) and validated using 983 samples (30%). The results revealed that slope, rainfall, stream density, and LST were significant predictors of forest fire occurrence, with deciduous and evergreen forests exhibiting the highest susceptibility among land-use classes. The resulting forest fire susceptibility map classified 235.12 km² (21.16%) and 204.16 km² (18.38%) of the district as very high and high susceptibility, respectively, primarily in mountainous forest areas. The model achieved an overall accuracy of 77.5% and an Area Under the Curve (AUC) value of 0.852, indicating good predictive performance. Furthermore, the proposed Geographic Information System-Logistic Regression (GIS-LR) framework provides an interpretable and transferable approach for forest fire susceptibility assessment and generates spatial information that can support forest fire prevention, resource allocation, and environmental management in Northern Thailand and other fire-prone regions. Full article

Concealed fires within the insulation layers of buildings, such as cold storage facilities and cinemas, present a serious fire hazard because heat generated by electrical faults can accumulate behind protective panels before ignition and then spread rapidly once combustion begins. Conventional fire detection methods have limited capability to identify these hidden thermal abnormalities at the pre-ignition stage. To address this problem, this paper proposes a deep learning method, called the Multi-Scale Cross-Modal Fusion Network (MSCMFNet), that uses multispectral images to identify abnormal heat sources beneath insulation layers before visible combustion occurs. A standardized experimental platform was developed to accurately simulate subsurface heat sources within the pre-ignition temperature range of insulation materials. Instead of relying on fixed temperature thresholds, the proposed method learns the characteristic spectral patterns produced by hidden heating. It extracts information from different spectral bands, combines these complementary features, and verifies the persistence of detected heat sources over time to reduce false alarms caused by non-fire disturbances. Experimental results demonstrate that the proposed method can effectively detect concealed thermal anomalies before ignition, providing reliable early warning and offering a promising approach to improving fire safety in buildings that make extensive use of insulation materials. Full article