Search Results (6,675)

Maize kernel damage detection is critical for grain quality assessment and post-harvest processing. However, existing deep learning methods struggle to balance accuracy, model complexity, and multi-class recognition under real-world conditions. To address these issues, this paper proposes YOLOv11n_DEG, an improved lightweight detection model based on YOLOv11n. The model uses the first ten pretrained layers as a feature extractor, replaces standard convolutions in the backbone with depthwise separable convolutions to reduce parameters, integrates an Efficient Channel Attention (ECA) module to enhance feature representation, and employs a dual-dropout strategy in the classification head to mitigate overfitting and improve generalization. Additionally, to account for potential discrepancies in damage characteristics between the obverse and reverse sides of maize kernels, a dual-sided synchronous image feature fusion method is introduced. The output layer classifies five target categories for multi-class damage detection. On an independent test set, the proposed model achieves a precision of 92.5%, a recall of 89.6%, and a mean average precision (mAP) of 93.4%, outperforming the original YOLOv11n by 6.2%, 4.0%, and 3.7%, respectively, while reducing parameter count and computational complexity by 32.9% and 6.5%. To validate practical deployability, a custom testbed with dual-camera synchronous acquisition and geometry-based matching was developed. On this platform, the model with dual-sided fusion achieves a single-side recognition accuracy of 94.1% and a dual-sided recognition accuracy of 89.3%, with an average detection time of 0.8 s per batch. These results demonstrate that YOLOv11n_DEG provides an accurate and practical solution for intelligent maize kernel damage detection, with strong potential for real-world deployment in grain inspection systems. Full article

During the fabrication of prestressed concrete beams, the quality and positional accuracy of the laid corrugated ducts (or prestressing ducts) directly influence the load-bearing capacity and durability of the beams. However, traditional manual inspection is inefficient, highly subjective, and difficult to achieve full coverage. To address this problem, this paper proposes an automated detection method that integrates improved YOLOv11-based pose estimation, robust curve fitting, and image stitching techniques. The method automatically identifies duct positions and evaluates laying quality. By incorporating the SE channel attention mechanism and the SPPFCSPC multi-scale pooling module, the SC-YOLOv11 model is developed, which significantly enhances the detection accuracy of slender corrugated pipe key points in environments with dense rebar occlusion. The RANSAC algorithm is employed to fit curves to the predicted key points, effectively suppressing the influence of outliers. Furthermore, the SIFT algorithm is used for precise stitching of drone-captured segmented images, which are then transformed into a unified front orthographic coordinate system of the entire box girder via perspective transformation, enabling accurate reconstruction of the corrected 2D layout of corrugated ducts across the full beam. Ablation experiments using 5-fold cross-validation demonstrate that SC-YOLOv11 improves mAP₅₀ and mAP_50–95 by 2.6% and 1.2%, respectively, with statistical significance (paired t-test, p < 0.01). The model achieves a per-image inference time of 6.37 ms, with 4.34 M parameters and 8.1 GFLOPs, meeting real-time requirements. In a 30 m prefabricated box girder field application, the measured section trajectory fitting curves of the corrugated ducts were compared with the design alignment, successfully identifying two abnormal locations where the laying deviation exceeded the allowable threshold. Cross-validation with on-site inspector records shows that over 92% of the measurement points agree within ±10 mm. This method achieves a fully automated analysis chain from key point detection and curve fitting to deviation quantification, providing an efficient, non-contact, and traceable intelligent tool for quality control of bridge prestressed systems. Full article

Lightweight real-time detectors must balance long-range contextual modeling, crossscale relational reasoning, and deployment efficiency. We present MambaACE-YOLO, an integration-oriented framework that separates computation into intra-scale encoding, cross-scale relational reasoning, sparse multi-stage feature distribution, and multiscale detection decoding. Its primary contribution is the coordinated integration of prior frequency-decoupled and state-space ideas with detection-specific interfaces, targetscale leave-one-source aggregation, and physically prunable feature distribution, rather than a new frequency-transform, state-space, or hypergraph mechanism class. Building on prior frequency-decoupled hybrid visual Mamba research, its detection-oriented DMobileMamba backbone applies bidirectional state-space scanning only to low-frequency components, while directional high-frequency correction and multi-kernel depthwise convolutions preserve boundaries and local shape. Compact Partial-Channel HyperACE (CP-HyperACE) models cross-scale higher-order relations in a semantic subspace and uses target-scale leave-one-source aggregation. Selective Additive FullPAD (SA-FullPAD) projects each cross-scale increment once and selects injection paths through static, physically prunable gates. On MS COCO 2017 val, the unpruned MambaACE-YOLO-N achieves 42.5 AP with 2.6 M parameters, and MambaACE-YOLO-S achieves 48.8 AP with 9.0 M parameters. Under a common documented RTX 5090 TensorRT FP16 setting, the unpruned Nano model records 1.19 ms network-forward latency at 42.5 AP, whereas its physically pruned counterpart retains 42.4 AP and records 1.02 ms. Accuracy and latency values are single-run or single-record point estimates without reported variance, and the 42.5-AP result uses the 600-epoch schedule without a matched 600-epoch YOLOv13-N control. We distinguish published cross-paper results from same-framework, same-device measurements and assess the individual design choices through controlled ablations and physical-pruning experiments. Full article

Anterior segment optical coherence tomography (AS-OCT) is essential for structural assessment of the anterior eye, yet automated landmark localization remains challenged by pervasive speckle noise, indistinct tissue interfaces, and labor-intensive manual annotation with notable inter-observer variability. This study presents NSE YOLO, an enhanced YOLOv11 framework for high-precision landmark localization in AS-OCT images after implantable collamer lens (ICL) implantation. It integrates a dual-branch NewConv module for multi-scale feature extraction, a dual-additive residual self-attention block (SABlock) to suppress background interference, and a Mamba-based EfficientViMBlock embedded in the C3k2 module to balance global contextual modeling and computational efficiency. Validated on 672 expert-annotated postoperative ICL images from 60 patients, NSE YOLO achieved an mAP@0.5 of 90.7% and mAP@0.5:0.95 of 85.1%, outperforming the YOLOv11 baseline by 5.2% and 6.6% with only 2.86 million parameters. Bland–Altman analysis showed negligible systematic bias and narrow limits of agreement for anterior chamber depth. For iridocorneal angle measurements, directional deviations and wider limits of agreement were observed, with performance approaching the level of inter-observer variability among human annotators. NSE YOLO enables automated quantification of anterior chamber depth and bilateral iridocorneal angles for post-ICL follow-up assessment, providing preliminary technical validation supporting further external and device-level evaluation. Full article

Manual visual diagnosis of apple diseases suffers from low efficiency, strong subjectivity and poor scalability for large commercial orchards. Existing research mainly targets diseases on single plant organs, whereas full-growth-cycle detection has to cope with extreme multi-scale differences among lesions. For instance, Valsa canker on tree trunks leads to extensive cortical necrosis, while early-stage anthracnose on fruits appears as tiny spots spanning only a few pixels. These significant scale gaps necessitate robust spatial feature aggregation and anti-noise ability to resist complex background interference. Aiming to achieve rapid and precise detection of diseases on multiple apple organs including leaves, fruits, trunks and branches, this work presents an enhanced YOLOv11 model equipped with the Information Integration Attention (IIA) module. The IIA module is embedded into the key fusion layers of the backbone and neck networks. It strengthens the extraction of fine-grained lesion features, recovers spatial location information via a bidirectional attention mechanism, and suppresses noise induced by uneven lighting and intricate backgrounds. To guarantee stable convergence on low-resource computing devices, a tailored training scheme is designed. Experimental results on a seven-category dataset with 7406 images demonstrate that YOLOv11-IIA reaches a precision of 0.763, a recall of 0.819, mAP@50 of 0.857 and mAP@50-95 of 0.699, which achieves clear performance improvements over the original YOLOv11 (mAP@50 improved from 0.485 to 0.857) and other attention-augmented detectors. The model operates stably on an NVIDIA GTX 1050 4GB GPU with an inference speed of 16 FPS for 640 × 640 input images; comprehensive quantitative computational metrics including parameter count, FLOPs and memory consumption will be fully measured in subsequent UAV deployment experiments. The proposed method provides a reliable technical reference for intelligent apple disease monitoring in smart orchard systems. 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

Rapid and reliable crack-based visual damage detection after earthquakes is crucial for safe and effective disaster response. Manual inspections are often slow and hazardous for engineers in unstable structures. This study proposes a quadrupedal robotic inspection system for rapid post-earthquake crack-based visual damage detection in reinforced concrete structures. A Unitree Go2 robot equipped with an Intel RealSense D435i RGB-D camera collected a dataset of 3255 annotated crack images from both field and public sources. The YOLOv8n model, trained and deployed on an NVIDIA Jetson AGX Xavier, demonstrated high detection performance in laboratory tests on reinforced concrete specimens, with precision, recall, and mAP@50 values all exceeding 85%. The system provides fast, accurate, and automated structural health assessments, reducing human risk and improving inspection efficiency in hazardous post-disaster environments. Future work will focus on expanding damage detection capabilities and real-world deployment. Full article

Multi-access edge computing (MEC) enables computation-intensive perception and decision-making tasks in vehicular networks to be offloaded to nearby edge servers. Existing approaches usually fix the artificial intelligence (AI) inference model, overlooking how model selection jointly affects latency, energy consumption, and service reliability. We propose a hierarchical model selection and control (HMSC) framework based on deep reinforcement learning (DRL) for MEC-assisted vehicular networks. The framework couples a vehicle-layer MAPPO component that provides a communication interface representation for subchannel assignment and energy accounting with a centralized MEC-layer soft actor-critic (SAC) agent that, under SDN orchestration, adaptively selects lightweight or high-fidelity AI models and allocates computational resources. Accordingly, the core contribution of this paper lies in MEC-side model-aware computation control under an explicitly defined subchannel-contention abstraction, rather than in physical-layer transmit-power optimization. Both layers are guided by a composite objective that integrates normalized end-to-end (E2E) latency, normalized energy consumption, and a deadline-violation penalty. Using a discrete-time simulation framework, HMSC reduces E2E latency compared with static inference and non-hierarchical DRL baselines and sustains a higher deadline satisfaction ratio (DSR) under constrained uplink throughput and varying traffic loads. The learned policy is load-aware, favoring high-fidelity inference under light load and lightweight inference under congestion; a post hoc analysis using YOLOv5-family accuracy reference further quantifies the inference-quality implications of this adaptive selection behavior. These results show that coordinated MEC-side control of AI model selection and computation, under a shared deadline-aware objective, provides a robust latency–energy trade-off for MEC-assisted vehicular networks. Full article

Domain shift between microscopy imaging domains poses a significant challenge for deploying deep learning-based cell detection models across different experimental setups. Manual annotation of new microscopy datasets remains resource-intensive and time-consuming. This study presents a detection-oriented Cycle-Consistent Generative Adversarial Network (CycleGAN)-based annotation transfer framework for adapting a labeled B16BL6 source domain to a HeLa target domain. Annotated B16BL6 melanoma microscopy images are translated into the visual appearance of HeLa microscopy data while retaining their original bounding-box annotations, enabling YOLOv8x detector training without full manual annotation of the target domain. Three YOLOv8x configurations were compared: a source-only bright B16BL6 baseline, an intensity-inverted dark B16BL6 baseline, and the proposed CycleGAN-translated B16BL6 → HeLa configuration. Performance was evaluated using standard object detection metrics on a manually annotated 100-frame HeLa target-domain subset, together with complementary unsupervised proxy metrics on the full unlabeled HeLa dataset. The CycleGAN-trained detector achieved the highest supervised target-domain performance, with a precision of 0.244, recall of 0.353, F1-score of 0.288, and mAP@0.50 of 0.186, compared with mAP@0.50 values of 0.027 and 0.009 for the bright and dark baselines, respectively. It also achieved the highest exploratory composite reliability score on the full HeLa sequence. These findings demonstrate that source-to-target image translation improves detector generalization under the investigated B16BL6 → HeLa phase-contrast microscopy domain shift, thereby reducing the need for extensive manual target-domain annotation. Full article

Accurate detection of small, rotated objects in practical scenarios such as remote sensing imagery and autonomous driving remains a critical challenge. Existing methods often struggle to balance detection accuracy and real-time performance, and face limitations in capturing subtle features, fusing multi-scale information, and regressing oriented bounding box angles. To address these issues, this paper proposes an Adaptive and Scalable YOLO model named AS-YOLOR (Adaptive and Scalable YOLO for Rotated object detection), based on the YOLOv8 baseline. The model introduces systematic innovations at three levels: feature enhancement, attention mechanism, and feature fusion. Firstly, a high-resolution detection head is introduced, combined with a Cross-layer Soft-weighted Feature Fusion (CSFF) strategy, to adaptively fuse shallow detail information with deep semantic features, thereby significantly improving the model’s ability to capture features of tiny objects. Secondly, a Hybrid Attention Module (HAM), enhanced with global context, is designed. This module integrates spatial attention in parallel with an improved Global Context Aware Block, effectively strengthening the model’s focus on and discriminative power for key target regions. Finally, an Adaptive Feature Fusion Path Network (AFFPN) is constructed. By incorporating learnable weights and scale-sequence modeling, it achieves superior modeling of the nonlinear relationships between multi-scale features. Experimental results on the DOTA-V1.0 public dataset demonstrate that the proposed AS-YOLOR model achieves 82.72% on the core evaluation metric mAP@50, representing a significant improvement of 3.22 percentage points over the baseline YOLOv8 model, and outperforming various mainstream rotated object detection methods. Comprehensive ablation studies validate the effectiveness of each improved component within the model. This research provides a solution with strong practical potential for achieving efficient and high-precision detection of small, rotated objects. Full article

Reliable perception of young apple fruits in natural orchards is a prerequisite for automated thinning and intelligent orchard management, yet remains difficult in real field conditions due to small fruit size, dense distribution, branch–leaf occlusion, background similarity, and severe illumination degradation at night. This study presents FruitDet, a lightweight multi-module detector designed for robust day–night young apple fruit detection in complex orchard environments. A field dataset was established in a high-density apple orchard in Aksu, Xinjiang, covering daylight and low-light night-time scenes with diverse occlusion, scale, and illumination variations. To improve detection robustness without sacrificing computational efficiency, FruitDet combines three complementary mechanisms: an inverted-bottleneck-based multi-scale feature enhancement module for preserving small-fruit details, a channel–spatial attention module for suppressing foliage and illumination interference, and a lightweight Transformer-based context module for modeling long-range dependencies between fruits and surrounding orchard structures. In daytime scenes, FruitDet achieved 91.904% precision, 77.557% recall, 83.254% mAP50, and 66.427% mAP50–95; in night-time scenes, it maintained 90.107% precision, 75.135% recall, 80.544% mAP50, and 64.719% mAP50–95. Compared with mainstream detectors including YOLOv5n, YOLOv8n, YOLO11n, YOLO26n, Faster R-CNN, RT-DETR, and RT-DETRv2, FruitDet consistently delivered higher accuracy across lighting conditions. Ablation, visualization, public-dataset testing, and edge-deployment experiments verified that the proposed modules jointly improve small-object representation, background discrimination, low-light robustness, and real-time applicability. With 2.960 M parameters, 3.726 G FLOPs, and approximately 180 FPS, FruitDet offers a practical and efficient visual perception approach for Young fruit monitoring was conducted under both daytime and night-time orchard conditions covered in this study. All-weather orchard monitoring and robotic young-fruit thinning. The shareable data are available Full article

Deploying object detection models on low-power embedded devices requires a careful balance between detection accuracy, runtime performance, and energy efficiency, especially for aerial images that contain small objects. We present a hardware-measured evaluation of Edge-TPU-compatible detectors on an MCU-class NXP i.MX RT1176 (bare-metal FreeRTOS, Cortex-M7) driving a Coral Edge TPU over an internal USB 2.0 link. We benchmark quantized YOLOv5-based detectors at input resolutions from 256 to 1024 px, reporting the accuracy (size-stratified AP_S on the VisDrone person split) together with the measured per-frame and per-pixel energy, sampled at 100 Hz with an automatic window and N ≥ 5 repetitions. We use the energy per pixel as a resolution-normalized view, not as a new metric, to show where the analytical model fails. The measured per-pixel energy departs from the compute (MAC)-based model because the inference is transfer-bound: it is dominated by the instruction stream and input activations re-sent over USB on every invoke, while the parameters stay cached on-chip. An Edge TPU clock sweep and a compiler-byte analysis agree independently that about 79% of each invocation is the USB transfer. This gives concrete design rules (a lower clock is more energy-efficient; ReLU is preferred to SiLU), and shows that the USB 2.0 ceiling is architecturally fundamental for the MCU class. Full article

Synthetic aperture radar (SAR)–optical image fusion and object detection are two closely related tasks in remote sensing. Fusion can provide richer texture and structural cues for downstream detection, while detection can, in turn, provide object-level location and semantic information to improve fusion. However, effectively integrating these two tasks within a unified training framework remains challenging. Their optimization objectives are inherently different: fusion emphasizes cross-modal information preservation and structural fidelity, whereas detection focuses more on discriminative target representation. As a result, direct joint training often leads to mutual interference rather than mutual reinforcement. In addition, most existing joint frameworks remain serial or unidirectional, limiting effective bidirectional knowledge transfer between fusion and detection. To address these issues, we propose PCFD-Net (Parallel Collaborative Fusion-Detection Network), which consists of a fusion branch, a detection branch, and a bidirectional interaction branch, and unifies fused image generation and oriented object detection within a single training framework through explicit bidirectional interaction. The fusion branch employs dual ResNet-50 encoders, a multi-scale attention fusion module, and a progressive decoder, while the detection branch is built on YOLOv8. The key component of the proposed framework is the bidirectional interaction branch. On the one hand, the multi-scale fused features generated by the fusion branch are injected into the detection backbone to enhance the exploitation of cross-modal intermediate representations. On the other hand, we develop CSMDE (Category Semantic–Modality Disentangled Embedding), which disentangles category-discriminative and modality-preference semantics to map detector category outputs into instance-level semantic embeddings. These embeddings, together with object locations, are further fed into a dual-discriminator mechanism to reversely constrain the fusion branch, thereby strengthening SAR-discriminative target preservation and optical background structure consistency. Experiments on the M4-SAR and OGSOD1.0 datasets demonstrate that PCFD-Net consistently outperforms representative fusion and detection methods, achieving superior fusion quality and stronger downstream detection performance. Full article

Reliable detection of surface defects in hoist wire ropes is essential for the safe operation of hoisting systems. Because of the stranded rope structure, early broken wires and local scars often appear as weak, small-scale features embedded in repetitive textures. Axial rope motion introduces directional blur during image acquisition, further weakening defect boundaries and increasing missed detections and localization errors. To address these challenges, we propose RMP-YOLOv8n, an axial motion prior-guided detector based on YOLOv8n. The method incorporates a rope motion prior-guided high-frequency enhancement module (RMP-HFEM), which uses axial edge and motion high-pass responses to guide shallow feature enhancement, together with a P2 branch that retains spatial detail for small defects. All models were trained exclusively on clear images from the training split, while paired axial motion-blur subsets generated exclusively from the held-out test images were used only for robustness evaluation. Experimental results on the wire-rope defect dataset show that, under severe axial motion blur, the mAP@50 of the clear-trained YOLOv8n decreases from 94.87 ± 0.12% to 48.31 ± 0.11%, whereas RMP-YOLOv8n retains 90.62 ± 0.08% mAP@50 and 87.28 ± 0.15% recall across three training seeds, demonstrating substantially stronger robustness to axial motion blur. The proposed method provides technical support for the stable deployment of online hoist wire-rope inspection systems under high-speed operation, motion blur, and weak small-defect features. Full article

Accurate tomato maturity detection is essential for harvesting decisions, quality grading, and postharvest handling in greenhouse production. However, leaf occlusion, fruit overlap, complex backgrounds, illumination variation, and subtle color differences between adjacent maturity stages limit real-time detection performance. To improve accuracy and deployment efficiency, this study proposes TMAF-YOLO, a lightweight detection model based on YOLOv8n. In this model, LGhostConv replaces selected convolutional structures to reduce redundant computation. A tomato maturity-aware aggregation fusion module, termed TMAF, is introduced to enhance color, texture, and local structural feature representation in the detection branches. MA-CB Focal Loss is used to improve learning under class imbalance and hard-sample conditions. Experimental results showed that TMAF-YOLO achieved Precision, Recall, mAP50, and mAP50-95 values of 0.889, 0.873, 0.954, and 0.776, respectively. The model contained 2.647 M parameters and required 7.3 GFLOPs, with an inference speed of 209.030 FPS. Compared with YOLOv8n and heavier detection models, TMAF-YOLO achieved higher detection accuracy with fewer parameters and lower computational cost. It also outperformed YOLOv11n and YOLOv12n in detection accuracy while maintaining real-time inference performance. These results indicate that the proposed model is suitable for real-time greenhouse tomato maturity detection and can support automated harvesting and grading. Full article