Search Results (318)

Synthetic aperture radar (SAR) ship detection is essential for maritime surveillance, yet coherent speckle can obscure tiny offshore vessels, while high-response coastal structures can produce target-like responses. Repeated downsampling further weakens sparse vessel signatures, whereas a separate high-resolution prediction head increases dense prediction costs. We propose a dual-domain detector that embeds spatial–frequency encoding, multiscale refinement, and high-resolution feature transfer within the backbone–neck–pyramid representation. Scharr-derived boundary cues are coupled with learned frequency context and remain discriminative as features are fused across scales, while sample-preserving space-to-depth transformation and global–local receptive-field modeling integrate shallow detail into semantically stronger $P_3$ features without introducing a separate $P_2$ prediction head. Coordinated FSEM, MSFE, and SOEP modules preserve fine vessel structure and reduce clutter-induced ambiguity throughout the detection hierarchy. Experiments on SSDD, HRSID, and RSDD-SAR evaluate detection accuracy, cross-dataset transfer, limited-data behavior, sensitivity to controlled perturbations, and edge deployment. Across 30 seed-matched SSDD runs, the method improves mAP${}_{50:95}$ over the baseline by $2.70\pm 0.34$ percentage points. It achieves 69.3% and 71.5% mAP${}_{50:95}$ on HRSID and RSDD. On the Atlas 200I platform, the compiled FP16 model retains 69.3% mAP${}_{50:95}$ while running at 101.8 frames/s. Full article

Infrared detection of small ship targets is crucial for maritime surveillance and early warning. However, low-brightness, low-pixel-count vessels are easily confused with sea clutter and sensor noise. Existing multi-scale networks can enhance local responses and aggregate contextual features, but dense cross-scale interactions propagate these clutter responses, and mask-level supervision cannot clearly distinguish the vessel from the target-like background. To address this issue, we propose a novel infrared small-ship semantic segmentation method called TALES-Net based on semantic distance alignment and local evidence selection, which consists of a Local Evidence Selection Module (LESM) and a Semantic Distance Alignment Loss (SDAL). LESM enhances center-to-surround contrast and selectively propagates reliable evidence through sparse cross-scale channel interactions and spatially gated decoding. SDAL attracts vessel features to a learnable target prototype while rejecting target-like background features. Our proposed TALES-Net method effectively enhances the ability to distinguish weak target vessels from clutter backgrounds, thereby improving its internal representation capabilities in infrared small target detection. In experiments on the IRS-Ship dataset and controlled ablation, TALES-Net effectively improves target segmentation probability while maintaining a low false alarm rate. Extensive experiments on the IRS-Ship dataset show that TALES-Net achieves pixel accuracy (pixAcc), foreground intersection over union (IoU), target detection probability (PD), and false alarm rate (FA) of 95.84%, 91.70%, 87.70%, and 0.42%, respectively. These results demonstrate that LESM and SDAL can effectively improve the sensitivity of weak target ships and suppress false alarms under complex sea clutter conditions. Full article

Semantic segmentation of maritime imagery is challenging due to the coexistence of large homogeneous regions, such as sea and land, and small, sparsely distributed ship targets, particularly in complex port environments. Further, strong background dominance and intricate boundary structures often result in missed detections and inaccurate ship delineation with conventional segmentation frameworks. To address these challenges, this paper proposes DGDE-CSFNet, a Difference-Guided Dual-Branch Cross-Fusion Network with Cross-Scale Feature Fusion specifically designed for sea–land–ship segmentation. The network employs a dual-branch encoder in which convolutional and Transformer backbones operate in parallel to capture local spatial structures and long-range contextual dependencies. To exploit their complementary behaviors, the encoding stage incorporates a Hierarchical Difference-Guided Attention (HDGA) module. By explicitly characterizing inter-branch inconsistencies, the module selectively enhances discriminative responses that are essential for accurate small-object perception. The decoder further adopts a Multi-Level Cross-Scale Fusion (MCSF) to aggregate multi-resolution features, enabling effective interaction between shallow spatial details and deep semantic representations. This design improves boundary localization and structural consistency while suppressing background interference and preserving subtle ship details. Experimental results on the ISDSD and HRSC2016-SL datasets demonstrate that our method achieves consistent performance gains over existing state-of-the-art methods. Full article

Reliable ship detection in synthetic aperture radar (SAR) imagery is challenged by speckles, coastal clutter, scale variation, dense berthing, weak target responses, and limited computing resources. SLA-YOLO is a lightweight YOLO11n detector that coordinates established mechanisms across the backbone, neck, and head. LESCNet combines ECA-enhanced depthwise-pointwise convolution with residual large-kernel refinement; C3k2-DAPE applies input-conditioned branch weighting and channel recalibration at selected fusion blocks; and SA-LSCD shares intermediate transformations while retaining scale-specific predictors. On fixed SSDD and HRSID splits, SLA-YOLO achieves mAP@0.5 values of 97.3% and 90.8% and mAP@0.5:0.95 values of 68.2% and 64.3%, respectively, with 1.70 million parameters and 3.6 GFLOPs. Relative to YOLO11n, parameters and FLOPs decrease by 34.1% and 42.9%, while mAP@0.5 increases by 4.4 and 5.2 percentage points, respectively. A complete 2³ ablation shows positive average main effects for all three components and context-dependent interactions. Across four no-fine-tuning transfer settings, SLA-YOLO improves mAP@0.5 by 3.5–6.3 points. Three-seed experiments yield 97.17 ± 0.31% and 90.87 ± 0.35% mAP@0.5 on SSDD and HRSID. Small-ship AP improves over YOLO11n by 10.1 ± 0.5 and 6.9 ± 0.7 points, with larger gains inshore. Network-only throughput reaches 66.8 and 63.4 FPS on an RTX 3090; embedded-device performance remains unverified. Full article

Ship detection in synthetic aperture radar (SAR) imagery remains challenging because near-shore clutter, coherent speckle noise, dense scattering responses, and large target-scale variations often obscure vessel boundaries and weaken small-ship signatures. Although single-stage detectors provide efficient inference, their predominantly local convolutional modeling and fixed multiscale fusion strategies are insufficient for capturing long-range sea-surface context and adaptively emphasizing discriminative ship responses. To address these limitations, this paper proposes SeaMamba, a frequency-stabilized selective state-space multiscale detector for SAR ship detection in complex maritime scenes. Specifically, a frequency-domain speckle prior is introduced to stabilize SAR inputs while preserving target localization cues. A bidirectional selective state-space modeling module is then used to propagate long-range contextual information with input-adaptive scanning. Furthermore, a gated pyramid reassembly module is designed to refine multiscale features before dense prediction. The proposed method is evaluated on the SAR Ship Detection Dataset (SSDD) and High-Resolution SAR Images Dataset (HRSID) under a unified five-fold cross-validation protocol. SeaMamba achieved mean average precision at an intersection-over-union threshold of 0.5 (mAP@0.5) values of $99.16\pm 0.11\%$ on SSDD and $93.74\pm 0.15\%$ on HRSID. Per-category evaluation, ablation studies, efficiency analysis, and Grad-CAM-based interpretability visualization further demonstrate that SeaMamba improves small-vessel detection, suppresses near-shore false responses, and maintains a practical accuracy-efficiency trade-off. Full article

Conventional manual microscopic examination for ship ballast-water microalgae detection suffers from cumbersome classification procedures, subjective viability assessment, and low counting efficiency, making it difficult to accomplish multi-task detection simultaneously. To address these issues, this study takes Tetraselmis chui and Phaeodactylum tricornutum as research objects and constructs a trinity detection model integrating species classification, viability discrimination, and automatic counting based on the YOLOv8m network, combining the color and morphological features of neutral red vital staining. A four-class microalgae annotated dataset was established, and the model was trained with an input resolution of 800 × 800 pixels. The optimal confidence threshold (0.383) was determined on the validation set by maximizing the F1-score and then applied to test-set evaluation. The experimental results demonstrate that on 156 independent test-set images, the proposed model achieves an average precision of 95.9% and mAP@0.5 of 96% across the four microalgae classes, with the highest precision of 97.3% for live Phaeodactylum tricornutum. Under the optimal confidence threshold, the average counting accuracy on the test set reaches 92.3%. Visualization tests confirm that the model can simultaneously accomplish species identification, viability determination, and cell counting in a single inference, providing a technical reference and algorithmic support for intelligent ballast-water microalgae monitoring. Full article

Marine oil spill monitoring plays a critical role in environmental protection and emergency response. Synthetic aperture radar (SAR) provides all-day, all-weather, and wide-area imaging capabilities and has become a primary sensor for operational oil spill surveillance. Nevertheless, SAR oil spill segmentation remains challenging. Real slicks are easily confused with look-alike dark formations; oil film boundaries are weak and fragmented under speckle noise; and small and sparse targets, such as ships, are easily overlooked amid the dominant sea surface background and complex coastal structures. To address these challenges, this paper proposes SDE-Net, an encoder–decoder network built on a ConvNeXtV2-Tiny backbone and a UPerNet-style decoder with three task-oriented modules. The Strip-Directional Local Enhancement (SDLE) module refines elongated low-contrast cues in the shallow lateral features of $C_2$ and $C_3$ while limiting the indiscriminate enhancement of speckle-contaminated responses. The Dynamic Scale Pyramid Context (DSPC) module redesigns pyramid-based context aggregation at the deepest stage by introducing an input-conditioned softmax scale gate that adaptively reweights the contributions of predefined contextual branches. The Edge-Aware Dynamic Gated FPN (EDG-FPN) decoder combines DySample-based feature alignment and dynamic gated fusion with a boundary-aware gate driven by an edge prediction branch, modulating high-level semantic propagation in boundary-sensitive regions without requiring additional edge annotations. On the Oil Spill Detection Dataset, SDE-Net achieves a 72.94% mIoU and 82.88% mDice, outperforming the UPerNet baseline equipped with the same ConvNeXtV2-Tiny backbone by 4.25 and 3.84 percentage points, respectively. The results demonstrate improved oil spill delineation and the preservation of sparse small-class structures under complex SAR sea surface conditions. Full article

Transferring unmanned surface vehicle (USV) algorithms from simulation to physical vessels is constrained by heterogeneous hardware interfaces, degraded wireless links, and ambiguous boundaries between human and autonomous control. This study designs and implements a ship–shore cooperative experimental platform comprising a shore control station, a portable control terminal, and an onboard system. The platform integrates multimodal sensing, private-radio and 4G/5G communication, an independent short-range remote-control (RC) path, and hardware arbitration. Small, safety-relevant commands are transmitted redundantly over the heterogeneous links using a shared application protocol with sequence-based deduplication. A two-dimensional control-authority model separates the authorized control source (shore, portable terminal, or short-range RC) from the active onboard behavior (path following, local collision avoidance, or safety protection) to organize fail-safe degradation and control transfer. Geometric light detection and ranging (LiDAR) obstacle detection and optimal reciprocal collision avoidance provide a representative onboard avoidance workflow. Full-scale vessel tests closed the loop from mission dispatch and command parsing to actuation and status feedback and demonstrated autonomous navigation, human takeover, and local collision avoidance. Across four water environments, the platform recorded 5.39 h of multimodal data over 18.26 km of valid trajectories. The results establish a physical testbed for ship–shore cooperative control, algorithm transfer, and multimodal data acquisition. Full article

Remote sensing ship detection is of great significance in marine traffic monitoring, port management, national defense security and other fields. Due to the small target scale, complex backgrounds and sparse distribution of ships in remote sensing images, existing real-time detectors still face challenges in localization accuracy and feature utilization efficiency. Taking RT-DETR-s as the baseline, this paper proposes two encoder-level improvements to boost the performance of ship detection in remote sensing imagery. On the attention mechanism side, an Edge-Aware Self-Attention (EASA) module is proposed as the principal methodological contribution. The Scharr edge map is injected into the attention matrix before softmax with a learnable intensity, providing a structural prior that enhances attention interactions between positions with strong edge responses, including ship contours, coastlines, and wake boundaries. The semantic distinction between ship-relevant and irrelevant edges is learned through the query-key projections in the attention mechanism itself, rather than through the edge bias. In terms of feature fusion, a Shallow-Guided Deep Fusion (SGDF) module is designed as an efficient integration strategy of existing techniques (CBAM attention and cross-scale fusion), with its novelty in the attention-first, downsampling-later pipeline that preserves fine-grained spatial details for small-object detection. It injects high-resolution small-object context from the backbone’s shallow layers into deep features, effectively mitigating the loss of spatial information of small ships in deep feature maps via spatial-channel attention. Finally, the existing WIoUv3 loss function is adopted and systematically compared with six other IoU variants to identify the most suitable regression loss for remote sensing ship detection. Experiments demonstrate that compared with RT-DETR-s, the proposed method achieves 2.8% higher mAP50–95 and 3.2% higher APsmall (APs) while maintaining real-time inference speed (37.6 FPS). Full article

Maritime object detection is essential for port-based surveillance, ship tracking, and maritime security. However, practical implementation encounters two primary challenges: significant class imbalance (IMO identification numbers represent only 1.2% of labelled objects) and spatial inconsistency (predicted IMO numbers often appear beyond ship boundaries). Standard detectors and conventional class-balancing strategies fail to adequately address these difficulties, resulting in a persistent mismatch between research validation and practical performance. We present a hybrid autoencoder–YOLO framework with variable spatial regularisation. To the best of our understanding, this is the first methodology to simultaneously address class imbalance and spatial inconsistency in maritime IMO detection via reconstruction-guided learning and differentiable spatial regularisation. The model uses a YOLOv8 encoder shared by both a detection head (designed for ships and IMO numbers), and an auxiliary reconstruction decoder (a SkipDecoder with U-Net-style skip connections). A unique spatial limitation loss provides the physical restriction that each IMO number must be within a ship’s bounding box, using only anticipated boxes. Training occurs in two phases: autoencoder pretraining on the marine dataset, followed by phased joint optimisation with a curriculum schedule for the spatial weighting. In a dataset of 297 annotated images (utilising five-fold cross-validation with a 47-image preserved test set), our comprehensive model achieved 50.1% IMO AP50-95, 97.8% IMO precision, and 94.6% IMO F1-score on the test set, beating the baseline YOLOv8s by +7.8 percentage points, +6.5 percentage points, and +5.2 percentage points, respectively. Ablation studies indicate that reconstruction instruction improves IMO AP50-95 by +4.5 percentage points, while spatial regularisation adds +3.3 percentage points. Although ship detection results in a small compromise (ship AP50-95 decreases from 77.0% to 55.6%), this appears to be practically acceptable given the essential role of IMO numbers as unique ship IDs. Significantly, inference speed improves by 20% (8.0 ms per image on an NVIDIA A100 GPU) relative to YOLOv8s (10.0 ms). Comparisons with leading detectors (RetinaNet, Faster R-CNN, DETR, EfficientDet), all initialised with standard COCO-pretrained backbones and fine-tuned on our dataset, reveal that none achieve an IMO AP50-95 exceeding 42.3%, highlighting the task’s challenge and the accuracy of our design. The proposed framework shows potential to address the gap in practical implementation through reconstruction-based feature learning accompanied by a specified geometric baseline. It is precise, accurate, and fast, aligns with specific physics, and shows promise for real-time maritime surveillance applications, though we acknowledge the need for additional thorough verification across many operational environments. Full article

Inshore synthetic aperture radar (SAR) ship detection remains challenging because strong coastal clutter, speckle noise, and degraded target responses frequently result in false alarms and missed detections of small vessels. Moreover, local scattering-related responses and high-level semantic information exhibit different characteristics across network stages, making it difficult for a unified feature-learning framework to fully exploit their complementarity. To address these issues, we propose a Scattering–Semantic Collaborative Network based on an asymmetric dual-branch DINO architecture, termed S-DINO. Specifically, a scattering-guided local feature reconstruction module establishes correlations between dispersed high-response regions and selected dominant response centers to improve the representation of fragmented target structures and enhance small-ship representation. Furthermore, a Semantic–Scattering Dual-Driven Query Injection strategy combines normalized feature-space response magnitude with semantic confidence to guide candidate reference-box initialization and reduce the selection bias caused by strong coastal activations. Experiments on the SSDD and HRSID inshore subsets demonstrate that, compared with the baseline DINO, S-DINO improves mAP@50 by 5.6 and 11.5 percentage points and F1-score by 11.4 and 8.6 percentage points, respectively. These results indicate the effectiveness of collaboratively exploiting scattering-related response cues and semantic information for ship detection in complex inshore environments. Full article

Ship detection in Synthetic Aperture Radar (SAR) imagery is essential for maritime surveillance and situational awareness. Despite the advances of deep learning for SAR ship detection, small-target detection is still hindered by severe feature degradation from repeated down-sampling and insufficiently discriminative representations under weak scattering and complex background clutter. To alleviate this dilemma, a Spatial Prior-guided feature enhancement and Multi-branch Complementary learning framework is proposed, termed SPMC, for small ship detection in SAR images. Specifically, a Spatial Prior-Guided Feature Enhancement (SPFE) module is designed to derive spatial attention priors for multi-scale features from ground-truth annotations, thereby emphasizing target-related responses and strengthening small-ship representations. Second, a Multi-branch Complementary Classification (MCC) module is developed, which introduces multiple auxiliary classification heads to learn complementary discriminative information from different classification perspectives. Furthermore, a dual-weighted complementary regularization strategy is proposed to encourages different classifiers to focus on hard samples, thereby improving the discriminative capability for small ships. Extensive experiments on the HRSID and LS-SSDD benchmarks validate the effectiveness of the proposed framework. For extremely small ships with very limited image coverage, SPMC improves the baseline YOLOv11 detector by 3.25%/0.99% in AP₅₀/AP_0.5:0.95 on LS-SSDD and by 1.28%/1.21% on HRSID, demonstrating its effectiveness in challenging SAR small ship detection. Full article

Continuous maritime surveillance increasingly relies on optical remote sensing, yet small ships are difficult to detect in low-resolution imagery because limited spatial resolution, complex sea backgrounds, and sensor noise suppress the fine-grained cues on which detectors depend, while routine acquisition of high-resolution imagery is prohibitively costly. To address this problem, a hybrid framework (RGT-YOLOv5Det) is proposed that couples transformer-based super-resolution (SR) with a lightweight object detector so that fine spatial detail is restored before detection. Methodologically, a paired benchmark (Ship-HRRSI/Ship-LRRSI) was constructed from the public TGRS-HRRSD dataset by standardising images to 800 × 800 pixels and generating 200 × 200 pixel counterparts via 4× bicubic down-sampling; three SR models (RGT, HAT, and Real-ESRGAN) and four detectors (YOLOv5s, YOLOv8s, YOLOv10s, and Faster R-CNN-MobileNetV3-Large-FPN) were fine-tuned and compared under identical training settings, and the best-performing components were integrated into the proposed two-stage pipeline. In the results, RGT delivered the best reconstruction quality (PSNR 22.038 dB, SSIM 0.3502) with the fewest parameters (13.37 M), YOLOv5s proved the most resolution-robust detector, and the integrated RGT-YOLOv5Det achieved mAP@0.5 of 0.947 and mAP@0.5:0.95 of 0.768 on low-resolution imagery, exceeding the best standalone detector score on each metric by 0.033 and 0.089, respectively. It is concluded that restoring structural detail prior to detection offers an accurate and acquisition cost-efficient alternative to high-resolution imaging, providing a practical route to reliable small ship detection in degraded optical remote sensing imagery. Full article

Object detection in visible nearshore surveillance imagery is of great importance for maritime safety, intelligent coastal monitoring, and water rescue applications. Nevertheless, reliable detection remains difficult because nearshore scenes often contain numerous small targets, cluttered wave patterns, shoreline textures, and substantial illumination variations. Moreover, existing public datasets mainly emphasize vessel detection and provide limited nearshore object categories. To solve these limitations, this study presents a large-scale visible nearshore dataset containing 20,934 images annotated with seven categories: pedestrian, sailor, swimmer, ship, boat, flotage, and seamark. The dataset is designed to support comprehensive evaluation and fair comparison of detection algorithms in complex nearshore environments. Based on the proposed benchmark, we conduct extensive evaluations of multiple mainstream object detectors and further develop a detection framework termed VN-DETR. The proposed model enhances both feature extraction and multi-scale feature fusion for nearshore scenarios. Specifically, a kernel selective attention based on WTConv (WKSA) module is designed to enlarge the receptive field and exploit contextual information in visible images, enabling more accurate object classification. In addition, a cross-layer feature selection and fusion (CFSF) module is introduced to perform feature matching, selection, and fusion across adjacent layers, enhancing the discriminability between foreground objects and complex nearshore backgrounds. This design effectively improves robustness against background noise such as wave reflections and shoreline textures. Extensive experiments on the constructed dataset demonstrate that VN-DETR consistently outperforms representative baseline methods and achieves superior detection performance, particularly for challenging small object categories. 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