Multi-sensor optimal weighted fusion incremental Kalman smoother

In practical applications, the system observation error is widespread. If the observation equation of the system has not been verified or corrected under certain environmental conditions,the unknown system errors and filtering errors will come into being.The incremental observation equation is derived, which can eliminate the unknown observation errors effectively. Furthermore, an incremental Kalman smoother is presented. Moreover, a weighted measurement fusion incremental Kalman smoother applying the globally optimal weighted measurement fusion algorithm is given.The simulation results show their effectiveness and feasibility.

Self-tuning weighted measurement fusion Kalman filter and its convergence

For multisensor systems,when the model parameters and the noise variances are unknown,the consistent fused estimators of the model parameters and noise variances are obtained,based on the system identification algorithm,correlation method and least squares fusion criterion.Substituting these consistent estimators into the optimal weighted measurement fusion Kalman filter,a self-tuning weighted measurement fusion Kalman filter is presented.Using the dynamic error system analysis （DESA） method,the convergence of the self-tuning weighted measurement fusion Kalman filter is proved,i.e.,the self-tuning Kalman filter converges to the corresponding optimal Kalman filter in a realization.Therefore,the self-tuning weighted measurement fusion Kalman filter has asymptotic global optimality.One simulation example for a 4-sensor target tracking system verifies its effectiveness.

ADAPTIVE FUSION ALGORITHMS BASED ON WEIGHTED LEAST SQUARE METHOD

Weighted fusion algorithms, which can be applied in the area of multi-sensor data fusion, are advanced based on weighted least square method. A weighted fusion algorithm, in which the relationship between weight coefficients and measurement noise is established, is proposed by giving attention to the correlation of measurement noise. Then a simplified weighted fusion algorithm is deduced on the assumption that measurement noise is uncorrelated. In addition, an algorithm, which can adjust the weight coefficients in the simplified algorithm by making estimations of measurement noise from measurements, is presented. It is proved by emulation and experiment that the precision performance of the multi-sensor system based on these algorithms is better than that of the multi-sensor system based on other algorithms.

Dehazing algorithm using adaptive dark channel fusion and sky compensation

Aiming at the inaccurate transmission estimation problem of dark channel prior image dehazing algorithm in the sudden change area of depth of field and sky area,a dehazing algorithm using adaptive dark channel fusion and sky compensation is proposed.Firstly,according to the characteristics of minimum filtering of large window scale and small window scale in the dark channel prior,the fused dark channel is obtained by weighted fusion of the approximate depth of field relationship,thus obtaining the primary transmission.Secondly,use the down-sampling to optimize the primary transmission combined with gray scale image of haze image by fast joint bilateral filtering,then restore the original image size by up-sampling,and the compensation of the Gaussian function is used in the sky area to obtain corrected transmission.Finally,the improved atmospheric light is combined with atmospheric scattering model to recover haze-free image.Experimental results show that the algorithm can recover a large amount of detailed information of the image,obtain high visibility,and effectively eliminate the halo effect.At the same time,it has a better recovery effect on bright areas such as the sky area.

APPLICATION OF FUZZY LOGIC IN WEIGHTED INFORMATION FUSION OF HAND GEOMETRY AND PALM PRINTS

Based on the theory of fuzzy logic, the method of obfuscating coefficient and reliability to fuse the information of hand geometry and palm prints for identity discrimination is proposed. The experiment proves that the method is useful and effective. Its identification rate is up to 90%, which is 20%-30% higher than that of using hand geometry or palm prints singly,thus it can be widely used in highly demanded security field, such as finance, entrance guard, etc.

A new information fusion white noise deconvolution estimator

The white noise deconvolution or input white noise estimation problem has important applications in oil seismic exploration, communication and signal processing. By the modern time series analysis method, based on the autoregressive moving average （ARMA） innovation model, a new information fusion white noise deconvolution estimator is presented for the general multisensor systems with different local dynamic models and correlated noises. It can handle the input white noise fused filtering, prediction and smoothing problems, and it is applicable to systems with colored measurement noises. It is locally optimal, and is globally suboptimal. The accuracy of the fuser is higher than that of each local white noise estimator. In order to compute the optimal weights, the formula computing the local estimation error cross-covariances is given. A Monte Carlo simulation example for the system with Bernoulli-Gaussian input white noise shows the effectiveness and performances.

A new PQ disturbances identification method based on combining neural network with least square weighted fusion algorithm

A new method for power quality (PQ) disturbances identification is brought forward based on combining a neural network with least square (LS) weighted fusion algorithm. The characteristic components of PQ disturbances are distilled through an improved phase-located loop (PLL) system at first, and then five child BP ANNs with different structures are trained and adopted to identify the PQ disturbances respectively. The combining neural network fuses the identification results of these child ANNs with LS weighted fusion algorithm, and identifies PQ disturbances with the fused result finally. Compared with a single neural network, the combining one with LS weighted fusion algorithm can identify the PQ disturbances correctly when noise is strong. However, a single neural network may fail in this case. Furthermore, the combining neural network is more reliable than a single neural network. The simulation results prove the conclusions above.

Application of fuzzy logic in weighted information of fusion fingerprints and palm prints

It is a developing job to distinguish identifications with information fusion of fingerprints and palm prints. It is also a very effective way to resolve the problem of low identification rate and low stability of single biology characteristic identification. Based on the theory of fuzzy logic theory, we bring out the method of obfuscating weigh coefficient and reliability to fuse the information of fingerprints and palm prints to realize high identification rate. The experiment proves the feasibility and effectiveness of this method and the identification rate can be more than 90%, which contributes useful experience to the research of identification using biology characteristics.

Self-supervised recalibration network for person re-identification

The attention mechanism can extract salient features in images,which has been proved to be effective in improving the performance of person re-identification(Re-ID).However,most of the existing attention modules have the following two shortcomings:On the one hand,they mostly use global average pooling to generate context descriptors,without highlighting the guiding role of salient information on descriptor generation,resulting in insufficient ability of the final generated attention mask representation;On the other hand,the design of most attention modules is complicated,which greatly increases the computational cost of the model.To solve these problems,this paper proposes an attention module called self-supervised recalibration(SR)block,which introduces both global and local information through adaptive weighted fusion to generate a more refined attention mask.In particular,a special"Squeeze-Excitation"(SE)unit is designed in the SR block to further process the generated intermediate masks,both for nonlinearizations of the features and for constraint of the resulting computation by controlling the number of channels.Furthermore,we combine the most commonly used Res Net-50 to construct the instantiation model of the SR block,and verify its effectiveness on multiple Re-ID datasets,especially the mean Average Precision(m AP)on the Occluded-Duke dataset exceeds the state-of-the-art(SOTA)algorithm by 4.49%.

YOLO-RLC:An Advanced Target-Detection Algorithm for Surface Defects of Printed Circuit Boards Based on YOLOv5

Printed circuit boards(PCBs)provide stable connections between electronic components.However,defective printed circuit boards may cause the entire equipment system to malfunction,resulting in incalculable losses.Therefore,it is crucial to detect defective printed circuit boards during the generation process.Traditional detection methods have low accuracy in detecting subtle defects in complex background environments.In order to improve the detection accuracy of surface defects on industrial printed circuit boards,this paper proposes a residual large kernel network based on YOLOv5(You Only Look Once version 5)for PCBs surface defect detection,called YOLO-RLC(You Only Look Once-Residual Large Kernel).Build a deep large kernel backbone to expand the effective field of view,capture global informationmore efficiently,and use 1×1 convolutions to balance the depth of the model,improving feature extraction efficiency through reparameterization methods.The neck network introduces a bidirectional weighted feature fusion network,combined with a brand-new noise filter and feature enhancement extractor,to eliminate noise information generated by information fusion and recalibrate information from different channels to improve the quality of deep features.Simplify the aspect ratio of the bounding box to alleviate the issue of specificity values.After training and testing on the PCB defect dataset,our method achieved an average accuracy of 97.3%(mAP50)after multiple experiments,which is 4.1%higher than YOLOv5-S,with an average accuracy of 97.6%and an Frames Per Second of 76.7.The comparative analysis also proves the superior performance and feasibility of YOLO-RLC in PCB defect detection.

Missing Value Imputation for Radar-Derived Time-Series Tracks of Aerial Targets Based on Improved Self-Attention-Based Network

The frequent missing values in radar-derived time-series tracks of aerial targets(RTT-AT)lead to significant challenges in subsequent data-driven tasks.However,the majority of imputation research focuses on random missing(RM)that differs significantly from common missing patterns of RTT-AT.The method for solving the RM may experience performance degradation or failure when applied to RTT-AT imputation.Conventional autoregressive deep learning methods are prone to error accumulation and long-term dependency loss.In this paper,a non-autoregressive imputation model that addresses the issue of missing value imputation for two common missing patterns in RTT-AT is proposed.Our model consists of two probabilistic sparse diagonal masking self-attention(PSDMSA)units and a weight fusion unit.It learns missing values by combining the representations outputted by the two units,aiming to minimize the difference between the missing values and their actual values.The PSDMSA units effectively capture temporal dependencies and attribute correlations between time steps,improving imputation quality.The weight fusion unit automatically updates the weights of the output representations from the two units to obtain a more accurate final representation.The experimental results indicate that,despite varying missing rates in the two missing patterns,our model consistently outperforms other methods in imputation performance and exhibits a low frequency of deviations in estimates for specific missing entries.Compared to the state-of-the-art autoregressive deep learning imputation model Bidirectional Recurrent Imputation for Time Series(BRITS),our proposed model reduces mean absolute error(MAE)by 31%~50%.Additionally,the model attains a training speed that is 4 to 8 times faster when compared to both BRITS and a standard Transformer model when trained on the same dataset.Finally,the findings from the ablation experiments demonstrate that the PSDMSA,the weight fusion unit,cascade network design,and imputation loss enhance imputation performance and confirm the efficacy of our design.

Segmentation of Head and Neck Tumors Using Dual PET/CT Imaging:Comparative Analysis of 2D,2.5D,and 3D Approaches Using UNet Transformer

The segmentation of head and neck(H&N)tumors in dual Positron Emission Tomography/Computed Tomogra-phy(PET/CT)imaging is a critical task in medical imaging,providing essential information for diagnosis,treatment planning,and outcome prediction.Motivated by the need for more accurate and robust segmentation methods,this study addresses key research gaps in the application of deep learning techniques to multimodal medical images.Specifically,it investigates the limitations of existing 2D and 3D models in capturing complex tumor structures and proposes an innovative 2.5D UNet Transformer model as a solution.The primary research questions guiding this study are:(1)How can the integration of convolutional neural networks(CNNs)and transformer networks enhance segmentation accuracy in dual PET/CT imaging?(2)What are the comparative advantages of 2D,2.5D,and 3D model configurations in this context?To answer these questions,we aimed to develop and evaluate advanced deep-learning models that leverage the strengths of both CNNs and transformers.Our proposed methodology involved a comprehensive preprocessing pipeline,including normalization,contrast enhancement,and resampling,followed by segmentation using 2D,2.5D,and 3D UNet Transformer models.The models were trained and tested on three diverse datasets:HeckTor2022,AutoPET2023,and SegRap2023.Performance was assessed using metrics such as Dice Similarity Coefficient,Jaccard Index,Average Surface Distance(ASD),and Relative Absolute Volume Difference(RAVD).The findings demonstrate that the 2.5D UNet Transformer model consistently outperformed the 2D and 3D models across most metrics,achieving the highest Dice and Jaccard values,indicating superior segmentation accuracy.For instance,on the HeckTor2022 dataset,the 2.5D model achieved a Dice score of 81.777 and a Jaccard index of 0.705,surpassing other model configurations.The 3D model showed strong boundary delineation performance but exhibited variability across datasets,while the 2D model,although effective,generally underperformed compared to its 2.5D and 3D counterparts.Compared to related literature,our study confirms the advantages of incorporating additional spatial context,as seen in the improved performance of the 2.5D model.This research fills a significant gap by providing a detailed comparative analysis of different model dimensions and their impact on H&N segmentation accuracy in dual PET/CT imaging.

Aircraft attitude estimation of MEMS sensor based on modified particle filter

The non-linearity problem of aircraft system could not be overcome by using the MEMS sensor only.In order to improve the accuracy of aerial vehicle attitude,an aircraft attitude estimation of the MEMS sensor based on modified particle filter is proposed.The aircraft attitude is optimized by the conjugate gradient method,and the drift error of gyroscope is reduced.Moreover,the particle weight is updated by the observed value to obtain an optimized state estimate.Finally,the conjugate gradient method and the modified particle filter are weightily combined to determine the optimal weighting factor.The attitude estimation is carried out with STM32 and MEMS sensor as the core to design system.The experimental results show that the static and dynamic attitude estimation performances of the aircraft are improved.The performances are well,the attitude data is relatively stable,and the tracking characteristics are better.Moreover,it has better robustness and stability.

Decoupled Wiener state fuser for descriptor systems

By the modem time series analysis method, based on the autoregressive moving average （ARMA） innovation models and white noise estimation theory, using the optimal fusion rule weighted by diagonal matrices, a distributed descriptor Wiener state fuser is presented by weighting the local Wiener state estimators for the linear discrete stochastic descriptor systems with multisensor. It realizes a decoupled fusion estimation for state components. In order to compute the optimal weights, the formulas of computing the cross-covariances among local estimation errors are presented based on cross-covariances among the local innovation processes, input white noise, and measurement white noises. It can handle the fused filtering, smoothing, and prediction problems in a unified framework. Its accuracy is higher than that of each local estimator. A Monte Carlo simulation example shows its effectiveness and correctness.

Sonar Image Target Detection for Underwater Communication System Based on Deep Neural Network

Target signal acquisition and detection based on sonar images is a challenging task due to the complex underwater environment.In order to solve the problem that some semantic information in sonar images is lost and model detection performance is degraded due to the complex imaging environment,we proposed a more effective and robust target detection framework based on deep learning,which can make full use of the acoustic shadow information in the forward-looking sonar images to assist underwater target detection.Firstly,the weighted box fusion method is adopted to generate a fusion box by weighted fusion of prediction boxes with high confidence,so as to obtain accurate acoustic shadow boxes.Further,the acoustic shadow box is cut down to get the feature map containing the acoustic shadow information,and then the acoustic shadow feature map and the target information feature map are adaptively fused to make full use of the acoustic shadow feature information.In addition,we introduce a threshold processing module to improve the attention of the model to important feature information.Through the underwater sonar dataset provided by Pengcheng Laboratory,the proposed method improved the average accuracy by 3.14%at the IoU threshold of 0.7,which is better than the current traditional target detection model.

Distributed cubature Kalman filter based on observation bootstrap sampling

Aiming at the adverse effect caused by observation noise on system state estimation precision,a novel distributed cubature Kalman filter(CKF) based on observation bootstrap sampling is proposed.Firstly,combining with the extraction and utilization of the latest observation information and the prior statistical information from observation noise modeling,an observation bootstrap sampling strategy is designed.The objective is to deal with the adverse influence of observation uncertainty by increasing observations information.Secondly,the strategy is dynamically introduced into the cubature Kalman filter,and the distributed fusion framework of filtering realization is constructed.Better filtering precision is obtained by promoting observation reliability without increasing the hardware cost of observation system.Theory analysis and simulation results show the proposed algorithm feasibility and effectiveness.

Mutual-information based weighted fusion for target tracking in underwater wireless sensor networks

Underwater wireless sensor networks(UWSNs)can provide a promising solution to underwater target tracking.Due to limited energy and bandwidth resources,only a small number of nodes are selected to track a target at each interval.Because all measurements are fused together to provide information in a fusion center,fusion weights of all selected nodes may affect the performance of target tracking.As far as we know,almost all existing tracking schemes neglect this problem.We study a weighted fusion scheme for target tracking in UWSNs.First,because the mutual information(MI)between a node’s measurement and the target state can quantify target information provided by the node,it is calculated to determine proper fusion weights.Second,we design a novel multi-sensor weighted particle filter(MSWPF)using fusion weights determined by MI.Third,we present a local node selection scheme based on posterior Cramer-Rao lower bound(PCRLB)to improve tracking efficiency.Finally,simulation results are presented to verify the performance improvement of our scheme with proper fusion weights.

Efficient phase-induced gabor cube selection and weighted fusion for hyperspectral image classification

Spectral-spatial Gabor filtering(GF),a robust feature extraction tool,has been widely investigated for hyperspectral image(HSI)classification.Recently,a new type of GF method,named phase-induced GF,which showed great potential for HSI feature extraction,was proposed.Although this new type of GF possibly better explores the frequency characteristics of HSIs,with a new parameter added,it generates a much larger amount of features,yielding redundancies and noises,and is therefore risky to severely deteriorate the efficiency and accuracy of classification.To tackle this problem,we fully exploit phase-induced Gabor features efficiently,proposing an efficient phase-induced Gabor cube selection and weighted fusion(EPCS-WF)method for HSI classification.Specifically,to eliminate the redundancies and noises,we first select the most representative Gabor cubes using a newly designed energy-based phase-induced Gabor cube selection(EPCS)algorithm before feeding them into classifiers.Then,a weighted fusion(WF)strategy is adopted to integrate the mutual information residing in different feature cubes to generate the final predictions.Our experimental results obtained on four well-known HSI datasets demonstrate that the EPCS-WF method,while only adopting four selected Gabor cubes for classification,delivers better performance as compared with other Gabor-based methods.The code of this work is available at https://github.com/cairlin5/EPCS-WF-hyperspectral-image-classification for the sake of reproducibility.

A Robust Face Recognition Method Using Multiple Features Fusion and Linear Regression

This paper presents a robust face recognition algorithm by using transform domain-based multiple feature fusion and lin- ear regression. Transform domain-based feature fusion can provide comprehensive face information for recognition, and decrease the effect of variations in illumination and pose. The holistic feature and local feature are extracted by discrete cosine transform and Gabor wavelet transform, respectively. Then the extracted holistic features and the local features are fused by weighted sum. The fused feature values are finally sent to linear regression classifier for recognition. The algorithm is evaluated on AR, ORL and Yale B face databases. Experiment results show that our proposed algo- rithm could be more robust than those single feature-based algo- rithms under pose and expression variations.

Boosting Multi-modal Ocular Recognition via Spatial Feature Reconstruction and Unsupervised Image Quality Estimation

In the daily application of an iris-recognition-at-a-distance(IAAD)system,many ocular images of low quality are acquired.As the iris part of these images is often not qualified for the recognition requirements,the more accessible periocular regions are a good complement for recognition.To further boost the performance of IAAD systems,a novel end-to-end framework for multi-modal ocular recognition is proposed.The proposed framework mainly consists of iris/periocular feature extraction and matching,unsupervised iris quality assessment,and a score-level adaptive weighted fusion strategy.First,ocular feature reconstruction(OFR)is proposed to sparsely reconstruct each probe image by high-quality gallery images based on proper feature maps.Next,a brand new unsupervised iris quality assessment method based on random multiscale embedding robustness is proposed.Different from the existing iris quality assess-ment methods,the quality of an iris image is measured by its robustness in the embedding space.At last,the fusion strategy exploits the iris quality score as the fusion weight to coalesce the complementary information from the iris and periocular regions.Extensive experi-mental results on ocular datasets prove that the proposed method is obviously better than unimodal biometrics,and the fusion strategy can significantly improve therecognition performance.