Search Results (1,315)

To address the insufficient capability of YOLO-series models in representing structural information for foreign objects with diverse scales and morphologies, an improved algorithm named SSFE-YOLO is proposed. First, the Space-to-Depth Convolution (SPDConv) is adopted into the backbone network to preserve edge and texture details in shallow features during downsampling, thereby maintaining the integrity of critical target structures at the feature generation stage. Second, an adaptive receptive field enhancement module (ARFE) is designed by introducing parallel feature branches with varying receptive fields. This module performs adaptive fusion to bolster the structural perception of the network towards polymorphic foreign objects. Furthermore, a distribution-feature stable compensation module (DFSC) is designed to suppress feature distribution shifts caused by illumination variations and noise interference through structural consistency enhancement and stable distribution constraints, which significantly improves the stability of feature representation in complex environments. Finally, a dual-dimension optimized loss function (D${}^2$-OL) is constructed to achieve differentiated supervision for samples of varying quality and balanced optimization for multi-scale target detection by modulating the supervisory weights of feature layers and filtering effective training samples. Experimental results on a self-built mine conveyor belt dataset demonstrate that the proposed method achieves an mAP@0.5 of 90.5% and an mAP@0.5:0.95 of 59.1%, consistently outperforming mainstream models such as YOLOv8, YOLOv11, and YOLOv13. Simulation results indicate that the proposed approach effectively enhances the detection accuracy and robustness of foreign objects in mining environments, showcasing substantial potential for engineering applications. Full article

Accurate detection of oil leakage spots is essential for oilfield safety and environmental protection. However, UAV-based inspection in onshore oilfields often suffers from complex illumination conditions, such as low light, backlighting, and mixed shadows, which simultaneously degrade image visibility and obscure leakage-sensitive features, thereby causing missed detection of minute and weak-texture oil leakage targets. Unlike generic low-light enhancement or object detection tasks, the core challenge of onshore UAV oil leakage inspection lies in preserving leakage-oriented fine cues during enhancement while improving the detector’s ability to distinguish leakage targets from highly confusing oilfield backgrounds. To address this task-specific challenge, we propose a collaborative low-light enhancement and detection framework that jointly optimizes leakage-detail-preserving enhancement and multi-scale interference-suppressed detection. Specifically, an improved Retinex-based enhancement network is designed by integrating multi-scale feature aggregation, NAFNet-based denoising, and a CBAM attention mechanism to enhance brightness while preserving leakage details. The enhanced images are then fed into an improved YOLOv11 detector, where an AC-FPN module is adopted to strengthen multi-scale feature fusion and suppress background interference. Experiments on UAV oilfield datasets demonstrate that the proposed method achieves a precision of 94.25% and a mean average precision (mAP) of 87.54%, outperforming existing approaches. The proposed framework provides an effective and robust solution for oil leakage spot detection under complex illumination. Full article

Machine learning methods are increasingly applied in daylight performance assessment due to their ability to model complex nonlinear relationships within large datasets while offering substantially faster predictions than conventional simulation workflows. Within this framework, deep learning architectures provide enhanced representational capability for capturing spatial and geometric dependencies. However, existing approaches often lack seamless integration with parametric design environments and offer limited interpretability regarding the influence of design parameters. This paper presents DayANN (Daylight Artificial Neural Network), a feedforward deep neural network developed within a structured Grasshopper-to-machine learning workflow for analyzing daylight performance in a parametrically defined office space. The method employs Climate Studio for Grasshopper to generate 288 simulation scenarios, forming the training dataset for the predictive model. The proposed framework enables automated data transfer, model training, and performance feedback within an iterative design–evaluation loop. In addition to predictive accuracy, SHAP-based interpretability is incorporated to quantify the contribution of individual daylighting parameters. The model achieved high accuracy, with R² values of 0.988 for Useful Daylight Illuminance (UDI) and 0.947 for Daylight Factor (DF), demonstrating that DayANN serves as a computationally efficient, transparent surrogate model suitable for early-stage architectural decision-making. Full article

Tea buds are the key raw material for high-quality tea production, and their accurate perception is essential for intelligent harvesting and quality-oriented management. However, tea bud detection in mountainous large-leaf tea plantations remains challenging because small, densely distributed targets are embedded in complex field environments, significantly limiting the stability and accuracy of existing detection methods. To address these challenges, this study proposes an improved tea bud detection model, termed YOLO-LAR, for mountainous large-leaf tea plantations in Yunnan Province, China, which is developed as an enhanced framework based on the YOLOv11 baseline. YOLO-LAR improves feature representation through multi-scale feature fusion, enabling more effective detection of densely distributed small tea buds. In addition, an optimized downsampling strategy is employed to preserve critical spatial information, and a context-enhanced feature aggregation mechanism is introduced to strengthen robustness under complex backgrounds and illumination variations. The results demonstrate that YOLO-LAR achieves precision, recall, mAP@0.50, and mAP@0.50:0.95 of 0.959, 0.908, 0.961, and 0.814, respectively, outperforming mainstream YOLO-based models, including YOLOv11n, YOLOv10n, and YOLOv8n. These results indicate that YOLO-LAR provides an effective and practical solution for accurate tea bud detection, offering strong technical support for intelligent harvesting and precision management in mountainous tea plantation environments. Full article

Visual–inertial odometry (VIO) often struggles with illumination variations, sparse visual features, and inertial drift in complex indoor settings, leading to scale uncertainties and accumulated errors. To address these issues, this paper proposes a new UWB–VIO initialization method combined with an enhanced Robust error-state Kalman filter (Robust ESKF) fusion technique for mobile robot localization. During initialization, common problems include scale drift and heading inconsistency. To solve these, a direction-consistent constrained initialization model is developed. By jointly optimizing the scale factor and yaw angle, this model ensures consistent alignment between the visual–inertial and ultra-wideband (UWB) coordinate frames. This approach removes the need for external calibration and independent coordinate transformation, which are typically required by traditional methods. In the fusion process, an improved residual-weighted robust filtering mechanism is employed to minimize the impact of abnormal UWB ranging data and noise interference. This mechanism adaptively suppresses outliers caused by UWB multipath reflections and non-line-of-sight (NLOS) propagation, thereby reducing VIO drift and improving the overall robustness and stability of the localization system. Experiments conducted in narrow-corridor environments, where both UWB and visual sensors are affected by interference, demonstrate that the proposed method significantly reduces trajectory drift and attitude jumps, resulting in better positioning accuracy and trajectory continuity. Compared to conventional UWB–VIO fusion algorithms, the proposed method enhances average localization accuracy by over 50% and maintains stable estimation even in severe multipath interference conditions, demonstrating high precision and strong robustness. Full article

Low-light image enhancement (LLIE) remains a challenging problem due to spatially varying illumination degradation, compressed tonal distributions, and structural detail loss. This paper presents Adaptive Multi-Branch Feature Fusion (AMBFF), a unified framework that formulates LLIE as a multi-domain representation alignment task. The proposed architecture explicitly models complementary feature domains, including hierarchical spatial context, luminance–chrominance decoupling, edge–texture structures, frequency-domain information, and differentiable tonal histogram representations. A spatially adaptive gating mechanism dynamically weights multi-feature branches through a convex fusion strategy, enabling location-aware illumination correction while preserving structural integrity and color fidelity. Extensive evaluations on widely used benchmark datasets demonstrate that AMBFF consistently outperforms representative conventional and deep learning-based approaches in terms of PSNR, SSIM, and LPIPS. Ablation analyses confirm the complementarity of the proposed feature domains and the robustness benefits of adaptive fusion. Despite its multi-branch design, AMBFF maintains a favorable performance–complexity trade-off, highlighting the effectiveness of structured multi-domain modeling for low-light image enhancement. Full article

In highway aerial imagery, complex weather conditions such as rain, fog, snow, and low illumination often lead to severe appearance degradation and feature loss of vehicle targets, posing significant challenges for vehicle detection. Existing research faces two major challenges: first, the lack of large-scale, high-quality annotated datasets tailored for complex weather scenarios; second, the difficulty traditional detectors encounter in effectively extracting feature information and performing multi-scale feature fusion under conditions of severe feature degradation and dense distribution of small objects. To address these issues, this paper investigates both data construction and algorithm design. Firstly, a Complex Weather Highway Vehicle Dataset (CWHVD) is established to provide a benchmark for related research. Secondly, a Feature-Enhanced Grid-Based Feature Fusion Complex-Weather Vehicle Detection Network (FGCV-Det) is proposed. A wavelet transform-based Feature Enhancement Module (FEWT) is introduced at the input stage to strengthen edge and texture representation. In the backbone, Adaptive Pinwheel Convolution (APConv) and a C3K2-HD module based on Hidden State Mixer-Based State Space Duality (HSM-SSD) are employed to enhance semantic modeling. Furthermore, a Complex Weather Grid Feature Pyramid Network (CWG-FPN) is designed to achieve weighted cross-scale fusion. The FGCV-Det significantly outperforms YOLO11s on CWHVD, achieving 63.4% precision, 48.6% recall, 51.7% mAP50, and 28.2% mAP50:95. It also generalizes well, reaching 47.1% and 49.6% mAP50 on VisDrone2019 and UAVDT, respectively, surpassing baseline and mainstream detectors, demonstrating strong robustness and generalization capability. Full article

To address the technical bottlenecks of strong background radiation interference and weak target signals in mid-wave infrared (MWIR) hyperspectral mineral detection over complex terrain, this paper proposes a “full-link background radiation suppression” methodological framework. A coupled illumination-terrain-atmosphere-sensor radiative transfer model is constructed to systematically quantify how multidimensional parameters—such as observation geometry, surface temperature, elevation, aerosol optical depth, and water vapor content—influence the target background radiation contrast. The findings reveal that daytime observation, lower surface temperature, higher altitude, dry atmosphere, and moderate solar and observation zenith angles are key factors for maximizing the signal-to-noise ratio. Comprehensive optimization analysis demonstrates that observations during midday in autumn and winter achieve optimal performance, with the target background relative contrast potentially enhanced by up to 6.29 times compared to unfavorable conditions such as summer nights. This work elucidates the physical mechanisms governing MWIR hyperspectral detection efficacy in complex scenarios, provides direct parameter-optimization strategies for intelligent mission planning of spaceborne imaging systems, and holds significant value for advancing mineral remote sensing from “passive acquisition” to “cognitive detection”. Full article

High-precision 3D reconstruction of non-cooperative space targets is a critical technology for on-orbit servicing (OOS) and situational awareness, driven by the growing number of OOS missions. However, traditional visual algorithms struggle to acquire accurate geometric information due to the unique high-dynamic-range lighting and strong specular reflections characteristic of the space environment. This paper proposes Space-Gaussian, a compact 3D Gaussian reconstruction method tailored for complex lighting environments. Built upon the 3D Gaussian Splatting (3DGS) framework, the method incorporates a physically based rendering pipeline and a microfacet bidirectional reflectance distribution function model. By decoupling geometric structure from material properties and utilizing deferred rendering, it effectively suppresses geometric artifacts and specular highlights arising from non-Lambertian surface reflections. Comparative experiments on a high-fidelity simulation dataset demonstrate that Space-Gaussian outperforms mainstream methods—including Neural Radiance Fields (NeRF), Instant-NGP, GaussianShader, and 3DGS—in geometric reconstruction accuracy, novel view synthesis quality, and real-time rendering. On our self-created dataset, our approach achieves a significant performance boost over existing 3DGS methods. The results highlight its potential for high-fidelity, real-time 3D perception on resource-constrained spacecraft platforms. Full article

With the rapid evolution of remote sensing earth observation technology, high-resolution object detection is crucial in military and civilian domains but faces challenges from expansive views and complex backgrounds. Small objects are particularly challenging due to their low pixel coverage, poor textures, and susceptibility to drastic illumination changes and background clutter. To address these problems, this paper proposes MDCA-YOLO for oriented object detection. A Dual-Branch Perception Module (DBPM) is designed utilizing a synergistic mechanism of large-kernel and strip convolutions to establish long-range dependencies, accurately capturing geometric features of tiny objects even in the absence of local details; Multi-Adaptive Selection Fusion (MASF) is proposed to address cross-scale feature loss by adaptively enhancing feature response while suppressing background noise; furthermore, a reconstructed decoupled detection head, CoordAttOBB, significantly improves angle regression accuracy while reducing complexity. Experimental results on the DIOR-R dataset show MDCA-YOLO surpasses YOLO11s, improving mAP50 and mAP50:95 by 2.5% and 2.7%, respectively, effectively proving the algorithm’s superiority in remote sensing tasks. Full article

The food industry increasingly relies on automated vision systems to ensure product quality, consistency, and safety. However, the visual classification of vegetables remains challenging due to high intra-class variability, illumination differences, and subtle morphological similarities between categories. This study evaluates the effectiveness of combining CNNs with four advanced Vision Transformer (ViT) architectures: DeiT (Data-efficient Image Transformer), CoaT (Co-Scale Conv-Attentional Transformer), CvT (Convolutional Vision Transformer), CrossViT (Cross-Attention Vision Transformer) for the automatic classification of 15 vegetable types. All models were implemented within a unified CNN–ViT hybrid framework to enhance both local feature extraction and global contextual reasoning. We processed all images under identical conditions to ensure a fair comparison and reproducibility. Results demonstrate that the hybrid architectures significantly outperform the standalone CNN baseline, with CvT achieving an approximate global accuracy in the range of 96.6–98.88% and consistently strong performance across visually complex classes such as cabbage, brinjal, and pumpkin. These findings confirm that hybrid CNN–ViT models are highly effective for visual food analysis, offering a robust and scalable solution for quality control, automated inspection, and classification of agricultural products. The methodology presented here may also be extended to other food items, including gels and processed products, highlighting its versatility and industrial relevance. Full article

Deep learning-based sheep face recognition technology significantly enhances the automation of individual sheep identification, providing critical technical support for smart livestock farming and precision agriculture. However, in real farming environments, factors such as complex backgrounds, illumination variations, and the high visual similarity of sheep faces severely constrain the comprehensive performance of recognition systems regarding accuracy and real-time capability. To address these challenges, we propose a cascaded framework comprising the WRT-DETR model for detection and LG-MobileViT for identification. WRT-DETR integrates multi-scale wavelet residual modeling and adaptive feature interaction into the RT-DETR architecture to effectively handle complex backgrounds. Subsequently, LG-MobileViT utilizes local–global collaborative modeling to distinguish fine-grained features while maintaining a lightweight footprint suitable for edge devices. Experiments conducted on a dataset of 400 individuals and 20,000 images demonstrate that WRT-DETR achieves 92.5% mAP50 in detection tasks. Furthermore, LG-MobileViT attains 98.97% recognition accuracy with a parameter size of only 4.57 MB. On edge computing platforms, the integrated system reaches an inference speed approaching 100 FPS. These results confirm that the proposed framework offers an efficient, reliable technical solution for non-contact, precise sheep identification in practical precision agriculture scenarios. Full article

Automotive vision systems depend critically on front-view cameras, whose image quality frequently degrades under adverse conditions such as rain, fog, low illumination, and rapid motion. To address this challenge, we propose VMESR, a variable mamba-enhanced super-resolution network that integrates a selective state-space model into a lightweight super-resolution architecture. By serializing 2D feature maps and applying variable-depth mamba blocks, VMESR captures long-range dependencies with linear complexity. A multi-scale feature extractor, enhanced residual modules equipped with a convolutional block attention module, and dense fusion connections work together to improve the recovery of high-frequency details. Extensive experiments demonstrate that VMESR achieves competitive performance in both objective metrics and perceptual quality compared to state-of-the-art methods, while significantly reducing parameter counts and computational cost. VMESR provides a practical balance between efficiency and reconstructive accuracy, offering a deployable super-resolution solution for embedded automotive sensors and enhancing the robustness of autonomous driving perception pipelines. Full article

Gastrointestinal cancers account for roughly a quarter of global cancer incidence, and early detection through endoscopy has proven effective in reducing mortality. Multi-class endoscopic disease detection, however, faces three persistent challenges: feature redundancy from non-pathological content, severe illumination inconsistency across imaging modalities, and extreme scale variability with blurry boundaries. This paper introduces Endo-DET, a domain-specific detection framework addressing these challenges through three synergistic components. The Adaptive Lesion-Discriminative Filtering (ALDF) module achieves lesion-focused attention via sparse simplex projection, reducing complexity from $O\left(N^2\right)$ to $O\left(\alpha N^2\right)$. The Global–Local Illumination Modulation Neck (GLIM-Neck) enables illumination-aware multi-scale fusion through four cooperative mechanisms, maintaining stable performance across white-light endoscopy, narrow-band imaging, and chromoendoscopy. The Lesion-aware Unified Calibration and Illumination-robust Discrimination (LUCID) module uses dual-stream reciprocal modulation to integrate boundary-sensitive textures with global semantics while suppressing instrument artifacts. Experiments on EDD2020, Kvasir-SEG, PolypGen2021, and CVC-ClinicDB show that Endo-DET improves mAP50-95 over the DEIM baseline by 5.8, 10.8, 4.1, and 10.1 percentage points respectively, with mAP75 gains of 6.1, 10.3, 6.8, and 9.3 points, and Recall50-95 improvements of 10.9, 12.1, 11.1, and 11.5 points. Running at 330 FPS with TensorRT FP16 optimization, Endo-DET achieves consistent cross-dataset improvements while maintaining real-time capability, providing a methodological foundation for clinical computer-aided diagnosis. Full article

Optically pumped magnetometers (OPMs) provide a non-cryogenic alternative to superconducting quantum interference devices (SQUIDs) for detecting weak biomagnetic fields. We report the design, construction, and characterization of a single-cell intrinsic OPM gradiometer. The gradiometer employs a rubidium-87 vapor cell in an orthogonal pump and probe beam configuration. The pump beam was split to illuminate two parallel sensing regions of the cell, separated by a baseline of 3 cm, with opposing circular polarization. A linearly polarized probe beam propagated through both regions and was captured by a balanced polarimeter whose output directly measured the spatial magnetic gradient. This prototype achieved a common-mode rejection ratio exceeding 50 dB and a sensitivity of 267 pT/cm/√Hz without passive magnetic shielding, using active ambient-field coils. As a proof of concept, we recorded preliminary cardiac-synchronous magnetic measurements using an optical pulse sensor for beat segmentation. After bandpass filtering and ensemble averaging, a cardiac-synchronous waveform was observed, consistent with cardiac timing. Unlike many multi-cell gradiometers that require complex calibration, modulation, and passive shielding, this single-cell design reduces cost and complexity. Full article