Search Results (25)

To enable the deployment of road distress detection models on resource-constrained embedded platforms, this paper presents a compression case study based on the LRDD-YOLOv8n detector designed for real-time 1080p video input. Layer-adaptive magnitude-based pruning (LAMP) was integrated with channel-wise knowledge distillation. First, LAMP performs structured pruning adaptive global sparsity allocation to reduce parameters and FLOPs. Then, a larger teacher model (LRDD-YOLOv8s) with high structural similarity guides the pruned student to recover feature representations. Compared to the baseline LRDD-YOLOv8n (64.4% mAP@0.5, 2.02 M parameters, 5.9G FLOPs, and 55.5 ms GPU inference time on Jetson TX2), our compressed model under a 1/1.4 target compression ratio achieves a mAP@0.5 of 65.1% (an slight accuracy increment of 0.7%), while reducing parameters by 36.1% (to 1.29 M) and FLOPs by 30.5% (to 4.1 G). Deployed on the BOXER-8120AI edge platform (Jetson TX2), the optimized model achieves an average inference time of 48.3 ms per frame (a 13.0% latency reduction compared to the baseline model). In addition, a 20 FPS detection rate was sustained under the 30 FPS maximum hardware acquisition limit of the industrial camera stream. Kinematic and geometric analysis validates that this processing rate utilizes 66.7% of all physically available frames and establishes a 95.4% consecutive frame-to-frame spatial overlap at typical inspection vehicle speeds (40–60 km/h). Full article

The effective cleaning of corn prior to storage is crucial for ensuring grain quality and safety. Traditional Convolutional Neural Network (CNN)-based detection methods often struggle to maintain accuracy in scenarios with dense occlusions. Furthermore, limitations in image quality and feature representation hinder their generalization to diverse impurity types. To address these challenges, this paper proposes an enhanced real-time detector transformer model named RT-DETR-CD (Real-Time Detector Transformer with Convolution and Dynamic Upsampling) for corn impurity detection based on industrial vision. This approach integrates Receptive Field Attention Convolutions (RFAConv) to enhance sensitivity to local texture details and employs the dynamic upsampling operator DySample to restore high-frequency edge information. Additionally, a novel Inner-Shape-IoU loss function is introduced to accelerate bounding box regression for objects with varying aspect ratios. Images were captured using FLIR industrial cameras under controllable annular LED illumination. Experiments on a self-built dataset demonstrate that the proposed model achieves a 4.7% improvement in mean average precision (mAP) and operates at 68 frames per second (FPS), outperforming the original RT-DETR model in both accuracy and speed. This work provides a practical solution for real-time, high-precision impurity detection on grain processing lines. Full article

Intrusion detectors are often evaluated using average metrics at unconstrained thresholds, yet deployments require explicit control over false alarms. We investigate zero-day (out-of-distribution, OOD) intrusion detection under a target-FPR calibrated protocol, where a threshold is set on benign validation traffic to satisfy a target false positive rate α and transferred, unchanged, to a seen-test and OOD-test. Using CICIDS2017-derived host-session nodes aggregated in 1 min and 5 min windows, we compare tabular baselines, message-passing GNNs on a rule-based graph, and employ a method that builds a k-nearest-neighbor similarity graph with lightweight feature pre-smoothing. Robustness is measured using the OOD violation ratio, percentile tail risk, and feasibility under explicit false-alarm budgets. Base-graph GNNs exhibit heavy-tailed false-alarm amplification under OOD shifts: at α = 0.001, the p95 violation ratio reaches 68.50 (1 m) and 67.95 (5 m). In contrast, the proposed method reduces p95 to 3.41 (1 m) and 1.15 (5 m) and improves budget feasibility. We further verify robustness beyond a single held-out family by evaluating additional unseen-family splits (e.g., DDoS and DDoS+DoS) under the same calibrated operating point. We also quantify deployment-oriented cost via edge-list size and practical parsing/loading time. These findings suggest that similarity-based graphs with light pre-smoothing improve deployability under distribution shifts. Full article

In the context of Industry 4.0, reliable automatic inspection of weld surface defects is critical for structural safety, yet current deep learning-based detectors struggle with the extreme scale variation and anisotropic shapes characteristic of weld flaws such as pores, cracks, and lack of fusion. Existing YOLO-family models, although effective on general-purpose datasets, often fail to robustly localize tiny defects and long, slender discontinuities while remaining lightweight enough for industrial edge deployment. A critical research gap lies in the lack of task-specific optimization for weld defects: standard attention mechanisms are isotropic and cannot capture linear defect continuity, while existing loss functions ignore scale disparity between tiny pores (area < 100 pixels²) and large incomplete fusion defects (area > 5000 pixels²), leading to unstable regression. Here, we propose a dual-optimized lightweight YOLOv11 framework tailored for weld defect detection that addresses both feature representation and bounding-box regression. Here, we propose a dual-optimized lightweight YOLOv11 framework tailored for weld defect detection that addresses both feature representation and bounding-box regression. First, we introduce WeldSimAM, an enhanced attention module that augments parameter-free SimAM with directional (horizontal/vertical) and channel-wise enhancement to better capture the directional texture of linear weld defects. Second, we develop an Enhanced Normalized Wasserstein Distance (EnNWD) loss, which incorporates scale-disparity penalties and relative-area-based weighting to mitigate sample imbalance and improve regression accuracy for tiny and large-aspect-ratio targets. Validated via 10-fold cross-validation on three datasets (self-built + two public), the method achieves 99.48% mAP@0.5 and 73.29% mAP@0.5:0.95, outperforming YOLOv11 by 0.13 and 3.76 percentage points (p < 0.01, two-tailed t-test), with 5.21 MB and 132 FPS on NVIDIA RTX 4090. It also surpasses non-YOLO SOTA methods (e.g., EfficientDet-Lite3) by 3.8–5.5 percentage points in mAP@0.5 (p < 0.05), offering a practical real-time solution for industrial inspection. Full article

An improved adaptive constant false alarm rate (CFAR) detector based on fuzzy theory is proposed to address the issue of poor detection performance and robustness of the variability index (VI) class CFAR detectors due to the misjudgment of the background environment and other reasons. The integration of the order statistic threshold adjustable detection algorithm (OSTA) into the adaptive CFAR detector has the potential to address the aforementioned issue. Therefore, in a clutter edge environment, the ratio of the means of the leading and lagging windows is calculated separately, and the differences between these mean ratios and predefined thresholds are used as inputs to the fuzzy inference machine, and the background clutter estimation of the OSTA is determined based on the fuzzy output, which can extend the range of values of the background clutter estimation, and improve the detection performance of the OSTA in this environment. The Kaigh–Lachenbruch quantile detection algorithm (KLQ) exhibits robust detection performance in multiple-target environments. Therefore, KLQ is used to detect targets in this environment, further improving the detection performance of the detector. The experimental results show that in multiple-target environments with an average misjudgment rate of 27.48%, the proposed detector increases the detection probability by 15.58% compared to the recently proposed variability index heterogeneous clutter estimate modified ordered statistics CFAR detector (VIHCEMOS-CFAR), and in a clutter edge environment, the false alarm rate of the proposed detector was reduced by approximately 89.64% compared to VIHCEMOS-CFAR. Additionally, the effectiveness of the proposed detector is also validated by real clutter data. It can be seen that the proposed adaptive CFAR detector has better robustness to the misjudgment of the background environment and better overall detection performance regardless of the environment. Full article

Spectral Computed Tomography (CT) is a versatile imaging technique widely utilized in industry, medicine, and scientific research. This technique allows us to observe the energy-dependent X-ray attenuation throughout an object by using Photon Counting Detector (PCD) technology. However, a major drawback of spectral CT is the increase in noise due to a lower achievable photon count when using more energy channels. This challenge often complicates quantitative material identification, which is a major application of the technology. In this study, we investigate the Noise2Inverse image denoising approach for noise removal in spectral computed tomography. Our unsupervised deep learning-based model uses a multi-dimensional U-Net paired with a block-based training approach modified for additional energy-channel regularization. We conducted experiments using two simulated spectral CT phantoms, each with a unique shape and material composition, and a real scan of a biological sample containing a characteristic K-edge. Measuring the peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) for the simulated data and the contrast-to-noise ratio (CNR) for the real-world data, our approach not only outperforms previously used methods—namely the unsupervised Low2High method and the total variation-constrained iterative reconstruction method—but also does not require complex parameter tuning. Full article

The resulting image obtained from an X-ray imaging system depends significantly on the characteristics of the detector. In particular, when an X-ray image is acquired by thinning the detector, a relatively large amount of noise inevitably occurs. In addition, when a thick detector is used to reduce noise in X-ray images, blurring increases and the ability to distinguish target areas deteriorates. In this study, we aimed to derive the optimal X-ray image quality by deriving the optimal noise reduction parameters based on the non-local means (NLM) algorithm. The detectors used were of two thicknesses (96 and 140 μm), and images were acquired based on the IEC 62220-1-1:2015 RQA-5 protocol. The optimal parameters were derived by calculating the edge preservation index and signal-to-noise ratio according to the sigma value of the NLM algorithm. As a result, a sigma value of the optimized NLM algorithm (0.01) was derived, and this algorithm was applied to a relatively thin X-ray detector system to obtain appropriate noise level and spatial resolution data. The no-reference-based blind/referenceless image spatial quality evaluator value, which analyzes the overall image quality, was best when using the proposed method. In conclusion, we propose an optimized NLM algorithm based on a new method that can overcome the noise amplification problem in thin X-ray detector systems and is expected to be applied in various photon imaging fields in the future. Full article

In the context of economic boom and climate change, monitoring the spatio-temporal dynamics of forest fragmentation induced by disturbances and understanding its corresponding associated factors are critical for developing informed forest management strategies. In this study, based on multi-temporal Landsat images acquired from 1999 to 2020, a SVM classifier was first applied to produce high-accuracy land cover maps in Xinyu City. Next, morphological spatial pattern analysis (MSPA) was implemented to characterize the spatio-temporal patterns of forest fragmentation by producing maps of seven fragmentation components, including the core, islet, perforation, edge, bridge, loop, and branch. Then, both natural and human factors responsible for the observed forest fragmentation dynamics were analyzed using the geo-detector model (GDM). The results showed that over the past two decades, Xinyu City experienced a process of significant forest area loss and exacerbating forest fragmentation. The forest area decreased from 1597.35 km² in 1999 to 1372.05 km² in 2020. The areal ratio of core patches decreased by 8.49%, and the areal ratio of edge patches increased by 5.98%. Spatially, the trend of forest fragmentation exhibited a progressive increase from the southern and northern regions towards the central and eastern areas. Large-scale forest core patches were primarily concentrated in the northwestern and southwestern regions, while smaller core patches were found in the eastern and central areas. Notably, human activities, such as distance from the roads and land use diversity, were identified as significantly associated with forest fragmentation. The interaction effect of these factors had a greater impact on forest fragmentation than their individual contributions. In conclusion, Xinyu City possesses the potential to further alleviate forest fragmentation by employing the regional differentiation development strategies: (1) intensive development in the northwest and southern regions; (2) high-density development in the western, northwestern, and southern regions, and (3) conservation development in the southwest, northeast, and east-central regions, thus aligning with the path of local social advancement. Full article

Polarimetric Synthetic Aperture Radar (PolSAR) data is inherently characterized by speckle noise, which significantly deteriorates certain aspects of the quality of the PolSAR data processing, including the polarimetric decomposition and target interpretation. With the rapid increase in PolSAR resolution, SAR images in complex natural and artificial scenes exhibit non-homogeneous characteristics, which creates an urgent demand for high-resolution PolSAR filters. To address these issues, a new adaptive PolSAR filter based on joint similarity measure criterion (JSMC) is proposed in this paper. Firstly, a scale-adaptive filtering window is established in order to preserve the texture structure based on a multi-directional ratio edge detector. Secondly, the JSMC is proposed in order to accurately select homogeneous pixels; it describes pixel similarity based on both space distance and polarimetric distance. Thirdly, the homogeneous pixels are filtered based on statistical averaging. Finally, the airborne and spaceborne real data experiment results validate the effectiveness of our proposed method. Compared with other filters, the filter proposed in this paper provides a better outcome for PolSAR data in speckle suppression, edge texture, and the preservation of polarimetric properties. Full article

Space vehicles such as missiles and aircraft have relatively long tracking distances. Infrared (IR) detectors are used for small target detection. The target presents point target characteristics, which lack contour, shape, and texture information. The high-brightness cloud edge and high noise have an impact on the detection of small targets because of the complex background of the sky and ground environment. Traditional template-based filtering and local contrast-based methods do not distinguish between different complex background environments, and their strategy is to unify small-target template detection or to use absolute contrast differences; so, it is easy to have a high false alarm rate. It is necessary to study the detection and tracking methods in complex backgrounds and low signal-to-clutter ratios (SCRs). We use the complexity difference as a prior condition for detection in the background of thick clouds and ground highlight buildings. Then, we use the spatial domain filtering and improved local contrast joint algorithm to obtain a significant area. We also provide a new definition of gradient uniformity through the improvement of the local gradient method, which could further enhance the target contrast. It is important to distinguish between small targets, highlighted background edges, and noise. Furthermore, the method can be used for parallel computing. Compared with the traditional space filtering algorithm or local contrast algorithm, the flexible fusion strategy can achieve the rapid detection of small targets with a higher signal-to-clutter ratio gain (SCRG) and background suppression factor (BSF). Full article

The performance of traditional model-based constant false-alarm ratio (CFAR) detection algorithms can suffer in complex environments, particularly in scenarios involving multiple targets (MT) and clutter edges (CE) due to an imprecise estimation of background noise power level. Furthermore, the fixed threshold mechanism that is commonly used in the single-input single-output neural network can result in performance degradation due to changes in the scene. To overcome these challenges and limitations, this paper proposes a novel approach, a single-input dual-output network detector (SIDOND) using data-driven deep neural networks (DNN). One output is used for signal property information (SPI)-based estimation of the detection sufficient statistic, while the other is utilized to establish a dynamic-intelligent threshold mechanism based on the threshold impact factor (TIF), where the TIF is a simplified description of the target and background environment information. Experimental results demonstrate that SIDOND is more robust and performs better than model-based and single-output network detectors. Moreover, the visual explanation technique is employed to explain the working of SIDOND. Full article

The non-uniformity of the readout circuit response in the infrared focal plane array unit detector can result in fixed pattern noise with stripe, which seriously affects the quality of the infrared images. Considering the problems of existing non-uniformity correction, such as the loss of image detail and edge blurring, a multi-scale residual network with attention mechanism is proposed for single infrared image stripe noise removal. A multi-scale feature representation module is designed to decompose the original image into varying scales to obtain more image information. The product of the direction structure similarity parameter and the Gaussian weighted Mahalanobis distance is used as the similarity metric; a channel spatial attention mechanism based on similarity (CSAS) ensures the extraction of a more discriminative channel and spatial feature. The method is employed to eliminate the stripe noise in the vertical and horizontal directions, respectively, while preserving the edge texture information of the image. The experimental results show that the proposed method outperforms four state-of-the-art methods by a large margin in terms of the qualitative and quantitative assessments. One hundred infrared images with different simulated noise intensities are applied to verify the performance of our method, and the result shows that the average peak signal-to-noise ratio and average structural similarity of the corrected image exceed 40.08 dB and 0.98, respectively. Full article

The paper focuses on the mathematical modeling of a new double linear array detector. The special feature of the detector is that image pairs can be generated at short intervals in one scan. After registration and removal of dynamic cloud edges in each image, the image differentiation-based change detection method in the temporal domain is proposed to combine with the structure tensor edge suppression method in the spatial domain. Finally, experiments are conducted, and our results are compared with theoretic analyses. It is found that a high signal-to-clutter ratio (SCR) of camera input is required to obtain an acceptable detection rate and false alarm rate in real scenes. Experimental results also show that the proposed cloud edge removal solution can be used to successfully detect targets with a very low false alarm rate and an acceptable detection rate. Full article

Recently, object detection in natural images has made a breakthrough, but it is still challenging in oriented ship detection for remote sensing imagery. Considering some limitations in this task, such as uncertain ship orientation, unspecific features for locating and classification in the complex optical environment, and multiplicative speckle interference of synthetic aperture radar (SAR), we propose an oriented ship detector based on the pairwise branch detection head and adaptive SAR feature enhancement. The details are as follows: (1) Firstly, the ships with arbitrary directions are described with a rotated ground truth, and an oriented region proposal network (ORPN) is designed to study the transformation from the horizontal region of interest to the rotated region of interest. The ORPN effectively improved the quality of the candidate area while only introducing a few parameters. (2) In view of the existing algorithms that tend to perform classification and regression prediction on the same output feature, this paper proposes a pairwise detection head (PBH) to design parallel branches to decouple classification and locating tasks, so that each branch can learn more task-specific features. (3) Inspired by the ratio-of-average detector in traditional SAR image processing, the SAR edge enhancement (SEE) module is proposed, which adaptively enhances edge pixels, and the threshold of the edge is learned by the channel-shared adaptive thresholds block. Experiments were carried out on both optical and SAR datasets. In the optical dataset, PBH combined with ORPN improved recall by 5.03%, and in the SAR dataset, the overall method achieved a maximum F1 score improvement of 6.07%; these results imply the validity of our method. Full article

Circle detection is a crucial problem in computer vision and pattern recognition. Improving the accuracy and efficiency of circle detectors has important scientific significance and excellent application value. In this paper, we propose a circle detection method with efficient arc extraction. In order to reduce edge redundancy and eliminate crossing points, we present an edge refinement algorithm to refine the edges into single-pixel-wide branchless contour curves. To address the contour curve segmentation difficulty, we improved the CTAR (Chord to Triangular Arms Ratio) corner detection method to enhance corner point detection and segment the contour curves based on corner points. Then, we used the relative position constraint of arcs to improve the circle detection accuracy further. Finally, we verified the feasibility and reliability of the proposed method by comparing our approach with five other methods using three datasets. The experimental results showed that the presented method had the advantages of anti-obscuration, anti-defect, and real-time performance over other methods. Full article