This article investigates the effects of blade damage on the dynamics of multirotor vehicles. A detailed explainable model that captures both the loss of effectiveness and the vibrations caused by damage to the propellers is proposed. Vibration contributions are calculated in the time domain and then further analyzed in the frequency domain, providing a theoretical justification for fault detection and isolation methods based on the frequency domain. The proposed model extends the classical control models adopted in the literature by including vibrations. The damping operated by the mechanical structure is finally estimated on an experimental case study, using flight data acquired from a hexarotor.
Drawing inspiration from the highly developed parallel perception and feedback loops of biological visual systems which excel at reconstructing physical continuity amidst intense clutter, this paper proposes a Visual Neuro Biomimetic Perception Network (VNBPNet) to resolve the structural fragmentation and perception-geometry decoupling inherent in Synthetic Aperture Radar (SAR) imaging. By establishing explicit closed-loop constraints between feature perception and geometric reconstruction, VNBPNet enables the simultaneous physical restoration of fragmented targets in low-contrast environments through three synergistic modules: the Neuro Visual Perception Module (NVPM), which leverages a retina-cortical antagonistic mechanism for adaptive clutter suppression and saliency enhancement; the Keypoint Reconstruction Module (KRM), which employs Gaussian diffusion guided by visual cortex completion mechanisms to restore physical continuity; and the Edge-Aware Module (EAM), which integrates edge details with high level semantics for precise boundary localization. Extensive evaluations on SSDD and HRSID benchmarks demonstrate that VNBPNet achieves competitive overall performance and obtains the highest mAP@0.5 and F1-score under the adopted evaluation setting. Ablation studies further confirm that this bio-inspired architecture yields significant gains over standard convolutional frameworks, particularly for multi-scale, low-contrast, and fragmented targets.
We present a distribution-theoretic framework for analyzing the limiting behavior of unit-mass function families under a flattening regime, leading to the formal notion of a unit uniform background ϵ(t). In contrast to the Dirac delta δ(t), obtained via concentration, this limit describes dispersion of mass over unbounded domains. Although the approximating sequences converge pointwise to zero while preserving unit mass, we show that they converge weakly to the zero distribution in S′(R), reflecting the non-commutation of limit and integration. To capture the asymptotic contribution of dispersed mass, we extend the test space to include bounded functions with well-defined limits at infinity. In this setting, the same sequences converge to an asymptotic evaluation functional U∞, yielding a hierarchy δ, ϵ, U∞, U that distinguishes concentration, dispersion, asymptotic evaluation, and global averaging. The proposed framework provides a unified interpretation of classical constructions, including improper priors in Bayesian inference and the distributional identity F[1]=2πδ(ω), and highlights the role of the test space in determining the limiting behavior of dispersive sequences.
To address the issue of low diagnostic model accuracy and generalization caused by state data distribution differences and sample imbalance across operating conditions and equipment, this paper proposes a cross-domain imbalanced sample fault diagnosis method based on parallel Temporal Convolutional Attention Network (TCAN) and Transformer, by integrating feature enhancement and transfer learning strategies. First, Refined Composite Hierarchical Fuzzy Dispersion Entropy (RCHFDE) is employed to extract features from state signals in both source and target domains, constructing an RCHFDE feature space to enhance state signals at the data level. Subsequently, a parallel TCAN-Transformer diagnostic model is designed and constructed, trained using complete, normal RCHFDE feature spaces from a specific operating condition or device. The trained parallel TCAN-Transformer diagnostic model is then transferred to new operating conditions or equipment. Using the RCHFDE feature space from limited unbalanced data of the new conditions or equipment, the diagnostic model undergoes parameter fine-tuning and training to obtain the final TL-TCAN-Transformer model suitable for the new context. This transferred diagnostic model is derived from the TCAN-Transformer model through parameter adjustment. Finally, the proposed method's effectiveness was validated using the CWRU bearing dataset and the research group's self-test bearing dataset. Experimental results demonstrate that the proposed method effectively improves cross-domain imbalance sample fault diagnosis accuracy and generalization capability.
This paper proposes a novel steganographic scheme for quantum remote sensing images. First, to enhance the invisibility and robustness of the secret information, an image encryption approach is developed by integrating quantum walks (QWs), a chaotic system, and the particle swarm optimization (PSO) algorithm. Subsequently, morphological edge detection is employed to extract and label comprehensive edge points within the quantum remote sensing images. Guided by this detection, a data hiding method based on a truncated octahedron in a unit cube (TO-UC) is designed for non-edge regions. This approach enables the highly effective concealment of the secret image while concurrently improving visual quality and security. Furthermore, an efficient and secure quantum steganographic circuit is constructed to facilitate its implementation. Experimental results demonstrate that the proposed scheme achieves a PSNR of up to 56.7797 dB and a SSIM index approaching 1. The secret image can be perfectly recovered in the absence of attacks; moreover, even under various levels of noise and clipping attacks, the NC and BER remain highly satisfactory. These results confirm the strong invisibility, high fidelity, and robust security of the proposed steganographic scheme.
Precise and automated ultrasound image segmentation is essential for improving computer-aided disease diagnosis. However, image degradations such as low contrast, intensity inhomogeneity, and speckle noise often obscure tissue details and reduce the accuracy of segmentation methods. While deformable models are theoretically explanatory and do not require large datasets, they are sensitive to initial contour placement. This article introduces a coarse-to-fine segmentation approach integrating deep learning with an adaptive coupled fractional deformable model (CFDM). Initially, the YOLOv9 (you only look once version 9) model is trained to generate coarse segmentations due to its boundary delineation efficiency. However, YOLOv9 struggles with degraded images and blurred edges, leading to coarse boundaries. To address this, the CFDM refines these boundaries by using adaptive fractional order, correcting inhomogeneity, and despeckling to improve segmentation accuracy. The well-posedness of the hybrid CFDM (HCFDM) is demonstrated, and experiments on breast ultrasound images of benign and malignant tumors show that the HCFDM outperforms five integer order deformable models, five fractional active contour models, four deep learning models, and two hybrid models. This makes it highly effective for automatic ultrasound image segmentation in clinical applications.
