Search Results (2,298)

The 3D reconstruction of point cloud trees and the acquisition of stand factors are key to supporting forestry regulation and urban planning. However, the two are usually independent modules in existing studies. In this work, we extended the AdTree method for 3D modeling of trees by adding a quantitative analysis capability to acquire stand factors. We used unmanned aircraft LiDAR (ALS) data as the raw data for this study. After denoising the data and segmenting the single trees, we obtained the single-tree samples needed for this study and produced our own single-tree sample dataset. The scanned tree point cloud was reconstructed in three dimensions in terms of geometry and topology, and important stand parameters in forestry were extracted. This improvement in the quantification of model parameters significantly improves the utility of the original point cloud tree reconstruction algorithm and increases its ability for quantitative analysis. The tree parameters obtained by this improved model were validated on 82 camphor pine trees sampled from the Northeast Forestry University forest. In a controlled experiment with the same field-measured parameters, the root mean square errors (RMSEs) and coefficients of determination (R²s) for diameters at breast height (DBHs) and crown widths (CWs) were 4.1 cm and 0.63, and 0.61 m and 0.74, and the RMSEs and coefficients of determination (R²s) for heights at tree height (THs) and crown base heights (CBHs) were 0.55 m and 0.85, and 1.02 m and 0.88, respectively. The overall effect of the canopy volume extracted based on the alpha shape is closest to the original point cloud and best estimated when alpha = 0.3. Full article

Gridless direction of arrival (DOA) estimation methods have garnered significant attention due to their ability to avoid grid mismatch errors, which can adversely affect the performance of high-resolution DOA estimation algorithms. However, most existing gridless methods are primarily restricted to applications involving uniform linear arrays or sparse linear arrays. In this paper, we derive the relationship between the element-domain covariance matrix and the angular-domain covariance matrix for arbitrary array geometries by expanding the steering vector using a Fourier series. Then, a deep neural network is designed to reconstruct the angular-domain covariance matrix from the sample covariance matrix and the gridless DOA estimation can be obtained by Root-MUSIC. Simulation results on arbitrary array geometries demonstrate that the proposed method outperforms existing methods like MUSIC, SPICE, and SBL in terms of resolution probability and DOA estimation accuracy, especially when the angular separation between targets is small. Additionally, the proposed method does not require any hyperparameter tuning, is robust to varying snapshot numbers, and has a lower computational complexity. Finally, real hydrophone data from the SWellEx-96 ocean experiment validates the effectiveness of the proposed method in practical underwater acoustic environments. Full article

In recent years, the non-invasive load monitoring (NILM) method based on sparse coding has shown promising research prospects. This type of method learns a sparse dictionary for each monitoring target device, and it expresses load decomposition as a problem of signal reconstruction using dictionaries and sparse vectors. The existing NILM methods based on sparse coding have problems such as inability to be applied to multi-state and time-varying devices, single-load characteristics, and poor recognition ability for similar devices in distributed manners. Using the analysis above, this paper focuses on devices with similar features in households and proposes a distributed non-invasive load monitoring method using Karhunen–Loeve (KL) feature extraction and an improved deep dictionary. Firstly, Karhunen–Loeve expansion (KLE) is used to perform subspace expansion on the power waveform of the target device, and a new load feature is extracted by combining singular value decomposition (SVD) dimensionality reduction. Afterwards, the states of all the target devices are modeled as super states, and an improved deep dictionary based on the distance separability measure function (DSM-DDL) is learned for each super state. Among them, the state transition probability matrix and observation probability matrix in the hidden Markov model (HMM) are introduced as the basis for selecting the dictionary order during load decomposition. The KL feature matrix of power observation values and improved depth dictionary are used to discriminate the current super state based on the minimum reconstruction error criterion. The test results based on the UK-DALE dataset show that the KL feature matrix can effectively reduce the load similarity of devices. Combined with DSM-DDL, KL has a certain information acquisition ability and acceptable computational complexity, which can effectively improve the load decomposition accuracy of similar devices, quickly and accurately estimating the working status and power demand of household appliances. Full article

This paper proposes a novel method capable of both detecting OOD data and generating in-distribution data samples. To achieve this, a VAE model is adopted and augmented with a memory module, providing capacities for identifying OOD data and synthesising new in-distribution samples. The proposed VAE is trained on normal data and the memory stores prototypical patterns of the normal data distribution. At test time, the input is encoded by the VAE encoder; this encoding is used as a query to retrieve related memory items, which are then integrated with the input encoding and passed to the decoder for reconstruction. Normal samples reconstruct well and yield low reconstruction errors, while OOD inputs produce high reconstruction errors as their encodings get replaced by retrieved normal patterns. Prior works use memory modules for OOD detection with autoencoders, but this method leverages a VAE architecture to enable generation abilities. Experiments conducted with CIFAR-10 and MNIST datasets show that the memory-augmented VAE consistently outperforms the baseline, particularly where OOD data resembles normal patterns. This notable improvement is due to the enhanced latent space representation provided by the VAE. Overall, the memory-equipped VAE framework excels in identifying OOD and generating creative examples effectively. Full article

Non-line-of-sight imaging is a technique for reconstructing scenes behind obstacles. We report a real-time passive non-line-of-sight (NLOS) imaging method for room-scale hidden scenes, which can be applied to smart home security monitoring sensing systems and indoor fast fuzzy navigation and positioning under the premise of protecting privacy. An unseen scene encoding enhancement network (USEEN) for hidden scene reconstruction is proposed, which is a convolutional neural network designed for NLOS imaging. The network is robust to ambient light interference conditions on diffuse reflective surfaces and maintains a fast reconstruction speed of 12.2 milliseconds per estimation. The consistency of the mean square error (MSE) is verified, and the peak signal-to-noise ratio (PSNR) values of 19.21 dB, 15.86 dB, and 13.62 dB are obtained for the training, validation, and test datasets, respectively. The average values of the structural similarity index (SSIM) are 0.83, 0.68, and 0.59, respectively, and are compared and discussed with the corresponding indicators of the other two models. The sensing system built using this method will show application potential in many fields that require accurate and real-time NLOS imaging, especially smart home security systems in room-scale scenes. Full article

The vertical distribution of the marine total suspended matter (TSM) concentration significantly influences marine material transport, sedimentation processes, and biogeochemical cycles. Traditional field observations are constrained by limited spatial and temporal coverage, necessitating the use of remote-sensing technology to comprehensively understand TSM variations over extensive areas and periods. This study proposes a remote-sensing approach to estimate the vertical distribution of TSM concentrations using MODIS satellite data, with the Bohai Sea and Yellow Sea (BSYS) as a case study. Extensive field measurements across various hydrological conditions and seasons enabled accurate reconstruction of in situ TSM vertical distributions from bio-optical parameters, including the attenuation coefficient, particle backscattering coefficient, particle size, and number concentration, achieving a determination coefficient of 0.90 and a mean absolute percentage error of 26.5%. In situ measurements revealed two distinct TSM vertical profile types (vertically uniform and increasing) and significant variation in TSM profiles in the BSYS. Using surface TSM concentrations, wind speed, and water depth, we developed and validated a remote-sensing approach to classify TSM vertical profile types, achieving an accuracy of 84.3%. Combining this classification with a layer-to-layer regression model, we successfully estimated TSM vertical profiles from MODIS observation. Long-term MODIS product analysis revealed significant spatiotemporal variations in TSM vertical distributions and column-integrated TSM concentrations, particularly in nearshore regions. These findings provide valuable insights for studying marine sedimentation and biological processes and offer a reference for the remote-sensing estimation of the TSM vertical distribution in other marine regions. Full article

Accurate monitoring of lithium-ion battery temperature is essential to ensure these batteries’ efficient and safe operation. This paper proposes a relevance-based reconstruction-oriented EMD-Informer machine learning model, which combines empirical mode decomposition (EMD) and the Informer framework to estimate the surface temperature of 18,650 lithium-ion batteries during charging and discharging processes under complex operating conditions. Initially, based on 9000 data points from the U.S. NASA Prognostics Center of Excellence’s random battery-usage dataset, where each data point includes three features: temperature, voltage, and current, EMD is used to decompose the temperature data into intrinsic mode functions (IMFs). Subsequently, the IMFs are reconstructed into low-, medium-, and high-correlation components based on their correlation with the original data. These components, along with voltage and current data, are fed into sub-models. Finally, the model captures the long-term dependencies among temperature, voltage, and current. The experimental results show that, in single-step prediction, the mean squared error, mean absolute error, and maximum absolute error of the model’s predictions are 0.00095, 0.02114, and 0.32164 °C; these metrics indicate the accurate prediction of the surface temperature of lithium-ion batteries. In multi-step predictions, when the prediction horizon is set to 12 steps, the model achieves a hit rate of 93.57% where the maximum absolute error is within 0.5 °C; under these conditions, the model combines high predictive accuracy with a broad predictive range, which is conducive to the effective prevention of thermal runaway in lithium-ion batteries. Full article

To improve the machining accuracy of free-form surface parts, a coarse-to-fine free-form surface machining error decomposition and compensation method is proposed in this paper. First, the machining error was coarsely decomposed using variational mode decomposition (VMD), and the correlation coefficients between the intrinsic mode function (IMF) and the machining error were obtained to filter out the IMF components that were larger than the thresholding value of the correlation coefficients, which was the coarse systematic error. Second, the coarse systematic errors were finely decomposed using empirical mode decomposition (EMD), which still filters out the IMF components that are larger than the thresholding value of the set correlation coefficient based on the correlation coefficient. Then, the wavelet thresholding method was utilized to finely decompose all the IMF components whose correlation coefficients in the first two decomposition processes were smaller than the threshold value of the correlation coefficient set. The decomposed residual systematic errors were reconstructed with the IMF components screened in the EMD fine decomposition, which gave the fine systematic error. Finally, the machining surface was reconstructed according to the fine systematic error, and its corresponding toolpath was generated to compensate for the machining error without moving the part. The simulation and analysis results of the design show that the method has a more ideal processing error decomposition ability and can decompose the systematic error contained in the processing error in a more complete way. The results of actual machining experiments show that, after using the method proposed in this paper to compensate for the machining error, the maximum absolute machining error decreased from 0.0580 mm to 0.0159 mm, which was a 72.5% reduction, and the average absolute machining error decreased from 0.0472 mm to 0.0059 mm, which was an 87.5% reduction. It was shown that the method was effective and feasible for free-form surface part machining error compensation. Full article

The parallel operation of multiple LCL-type converters will result in a deviation of the resonant frequency and resonance phenomena. The occurrence of harmonic resonance can cause problems such as an increase in harmonic voltage and current. This can lead to the malfunction of relay protection and automatic devices, causing damage to system equipment. In severe cases, it can cause accidents and threaten the safe operation of the power system. A hybrid damping active disturbance rejection control (HD-ADRC) method is proposed in this paper to suppress the harmonic resonance of parallel LCL-type converters. First, a third-order linear disturbance rejection controller (LADRC) including the linear extended-state observer and the error-feedback control rate is designed based on LCL-type converter model analysis. The proposed method considers the resonance couplings caused by both internal and external disturbances as the total disturbance, thus improving the anti-disturbance capabilities as well as the operational stability of converters in parallel. Then, a hybrid damping control is proposed to reconstruct the damping characteristics of converters to suppress the parallel resonance spike and reduce the resonance frequency offset. And the parameter selection of the control system is optimized through a stability analysis of the tracking performance and anti-disturbance performance of the HD-ADRC controller. Finally, all the theoretical considerations are verified by simulation and experimental results based on the Matlab/Simulink 2018B and dSpace platform. The simulation and experimental results show that the PI controller gives a THD of 5.33%, which is reduced to 4.66% by employing the HD-LADRC, indicating an improved decoupling between the converters working in parallel with the proposed control scheme. Full article

An adaptive neural network output-feedback control strategy is proposed in this paper for the distributed optimization problem (DOP) of high-order nonlinear stochastic multi-agent systems (MASs) driven by Lévy noise. On the basis of the penalty-function method, the consensus constraint is removed and the global objective function (GOF) is reconstructed. The stability of the system is analyzed by combining the generalized Itô’s formula with the Lyapunov function method. Moreover, the command filtering mechanism is introduced to solve the “complexity explosion” problem in the process of designing virtual controller, and the filter errors are compensated by introducing compensating signals. The proposed algorithm has been proved that the outputs of all agents converge to the optimal solution of the DOP with bounded errors. The simulation results demonstrate the effectiveness of the proposed approach. Full article

Thin clouds in Remote Sensing (RS) imagery can negatively impact subsequent applications. Current Deep Learning (DL) approaches often prioritize information recovery in cloud-covered areas but may not adequately preserve information in cloud-free regions, leading to color distortion, detail loss, and visual artifacts. This study proposes a Sparse Transformer-based Generative Adversarial Network (SpT-GAN) to solve these problems. First, a global enhancement feature extraction module is added to the generator’s top layer to enhance the model’s ability to preserve ground feature information in cloud-free areas. Then, the processed feature map is reconstructed using the sparse transformer-based encoder and decoder with an adaptive threshold filtering mechanism to ensure sparsity. This mechanism enables that the model preserves robust long-range modeling capabilities while disregarding irrelevant details. In addition, inverted residual Fourier transformation blocks are added at each level of the structure to filter redundant information and enhance the quality of the generated cloud-free images. Finally, a composite loss function is created to minimize error in the generated images, resulting in improved resolution and color fidelity. SpT-GAN achieves outstanding results in removing clouds both quantitatively and visually, with Structural Similarity Index (SSIM) values of 98.06% and 92.19% and Peak Signal-to-Noise Ratio (PSNR) values of 36.19 dB and 30.53 dB on the RICE1 and T-Cloud datasets, respectively. On the T-Cloud dataset, especially with more complex cloud components, the superior ability of SpT-GAN to restore ground details is more evident. Full article

Rubber (Hevea brasiliensis Muell.) plantations are vital agricultural ecosystems in tropical regions. These plantations provide key industrial raw materials and sequester large amounts of carbon dioxide, playing a vital role in the global carbon cycle. Climate change has intensified droughts in Southeast Asia, negatively affecting rubber plantation growth. Limited in situ observations and short monitoring periods hinder accurate assessment of drought impacts on the gross primary productivity (GPP) of rubber plantations. This study used GPP data from flux observations at four rubber plantation sites in China and Thailand, along with solar-induced chlorophyll fluorescence (SIF), enhanced vegetation index (EVI), normalized difference vegetation index (NDVI), near-infrared reflectance of vegetation (NIRv), and photosynthetically active radiation (PAR) indices, to develop a robust GPP estimation model. The model reconstructed eight-day interval GPP data from 2001 to 2020 for the four sites. Finally, the study analyzed the seasonal drought impacts on GPP in these four regions. The results indicate that the GPP prediction model developed using SIF, EVI, NDVI, NIRv, and PAR has high accuracy and robustness. The model’s predictions have a relative root mean square error (rRMSE) of 0.22 compared to flux-observed GPP, with smaller errors in annual GPP predictions than the MOD17A3HGF model, thereby better reflecting the interannual variability in the GPP of rubber plantations. Drought significantly affects rubber plantation GPP, with impacts varying by region and season. In China and northern Thailand (NR site), short-term (3 months) and long-term (12 months) droughts during cool and warm dry seasons cause GPP declines of 4% to 29%. Other influencing factors may alleviate or offset GPP reductions caused by drought. During the rainy season across all four regions and the cool dry season with adequate rainfall in southern Thailand (SR site), mild droughts have negligible effects on GPP and may even slightly increase GPP values due to enhanced PAR. Overall, the study shows that drought significantly impacts rubber the GPP of rubber plantations, with effects varying by region and season. When assessing drought’s impact on rubber plantation GPP or carbon sequestration, it is essential to consider differences in drought thresholds within the climatic context. Full article

A fast and accurate radiative transfer model is the prerequisite in the field of atmospheric remote sensing for limb atmospheric inversion to tackle the drawback of slow calculation speed of traditional atmospheric radiative transfer models. This paper established a fast computing model (ANN-HASFCM) for high-spectral-resolution absorption spectra by using artificial neural networks and PCA (principal component analysis) spectral reconstruction technology. This paper chose the line-by-line radiative transfer model (LBLRTM) as the comparative model and simulated training spectral data in the oxygen A-band (12,900–13,200 cm⁻¹). Subsequently, ANN-HASFCM was applied to the retrieval of the atmospheric density profile with the data of the Global Ozone Monitoring by an Occultation of Stars (GOMOS) instrument. The results show that the relative error between the optical depth spectra calculated by LBLRTM and ANN-HASFCM is within 0.03–0.65%. In the process of using the global-fitting algorithm to invert GOMOS-measured atmospheric samples, the inversion results using Fast-LBLRTM and ANN-HASFCM as forward models are consistent, and the retrieval speed of ANN-HASFCM is more than 200 times faster than that of Fast-LBLRTM (reduced from 226.7 s to 0.834 s). The analysis shows the brilliant application prospects of ANN-HASFCM in limb remote sensing. Full article

This paper presents an enhanced method for the transmission of 3D video in the Multi-view Video plus Depth (MVD) format over Two-Way Relay Channels (TWRC). Our approach addresses the unique challenges of MVD-based 3D video by combining Hierarchical Quadrature Amplitude Modulation (HQAM), a method that prioritizes data layers based on importance, and Inter-Layer Forward Error Correction (IL-FEC), which protects critical data from errors. These are specifically designed to handle the dual-layer data structure where color data and depth information require different levels of error protection, and it reduces transmission errors and enhances the quality of MVD-based 3D video over TWRC. In the TWRC scenario, the proposed scheme optimizes transmission by reducing the number of relayed bitstreams by half while maintaining high-quality requirements, as demonstrated by significant improvements in the Structural Similarity Index (SSIM) for virtually synthesized views. Furthermore, we identify and optimize the hierarchical modulation parameter ($\alpha$), which controls the priority and protection levels of different data streams. Systematically varying $\alpha$ reveals its substantial impact on the quality of the reconstructed 3D video, as measured by SSIM. Our results demonstrate that the proposed combination of HQAM and IL-FEC not only maintains the target SSIM of 0.9 for the virtually synthesized view under various relay conditions but also reveals the optimal $\alpha$ value for balancing the error protection between the color and depth map data streams. Notably, while increasing $\alpha$ enhances the protection of critical data (such as color video streams), it may concurrently degrade the quality of less important streams (like depth maps), highlighting the importance of fine-tuning $\alpha$ to achieve the best overall video quality. These findings suggest that our method provides a flexible and effective solution for high-quality 3D video transmission in challenging communication environments, potentially advancing the development of future 3D video delivery systems. Full article

This paper is concerned with an end-to-end electrocardiogram (ECG) signal transformation from a continuous-wave (CW) radar signal using a specialized deep learning model. For this purpose, the presented deep learning model is designed using convolutional neural networks (CNNs) and bidirectional long short-term memory (Bi-LSTM) with a maximum-overlap discrete wavelet transform (MODWT) layer and an adaptive neuro-fuzzy network (ANFN) layer. The proposed method has the advantage of developing existing deep networks and machine learning to reconstruct signals through CW radars to acquire ECG biological information in a non-contact manner. The fully connected (FC) layer of the CNN is replaced by an ANFN layer suitable for resolving black boxes and handling complex nonlinear data. The MODWT layer is activated via discrete wavelet transform frequency decomposition with maximum-overlap to extract ECG-related frequency components from radar signals to generate essential information. In order to evaluate the performance of the proposed model, we use a dataset of clinically recorded vital signs with a synchronized reference sensor signal measured simultaneously. As a result of the experiment, the performance is evaluated by the mean squared error (MSE) between the measured and reconstructed ECG signals. The experimental results reveal that the proposed model shows good performance in comparison to the existing deep learning model. From the performance comparison, we confirm that the ANFN layer preserves the nonlinearity of information received from the model by replacing the fully connected layer used in the conventional deep learning model. Full article