A Heterogeneous Information Fusion Method for Maritime Radar and AIS Based on D-S Evidence Theory

Maritime radar and automatic identification systems (AIS), which are essential auxiliary equipment for navigation safety in the shipping industry, have played significant roles in maritime safety supervision. However, in practical applications, the information obtained by a single device is limited, and it is necessary to integrate the information of maritime radar and AIS messages to achieve better recognition effects. In this study, the D-S evidence theory is used to fusion the two kinds of heterogeneous information: maritime radar images and AIS messages. Firstly, the radar image and AIS message are processed to get the targets of interest in the same coordinate system. Then, the coordinate position and heading of targets are chosen as the indicators for judging target similarity. Finally, a piece of D-S evidence theory based on the information fusion method is proposed to match the radar target and the AIS target of the same ship. Particularly, the effectiveness of the proposed method has been validated and evaluated through several experiments, which proves that such a method is practical in maritime safety supervision.

Hierarchical Visualized Multi-level Information Fusion for Big Data of Digital Image

At present,the process of digital image information fusion has the problems of low data cleaning unaccuracy and more repeated data omission,resulting in the unideal information fusion.In this regard,a visualized multicomponent information fusion method for big data based on radar map is proposed in this paper.The data model of perceptual digital image is constructed by using the linear regression analysis method.The ID tag of the collected image data as Transactin Identification(TID)is compared.If the TID of two data is the same,the repeated data detection is carried out.After the test,the data set is processed many times in accordance with the method process to improve the precision of data cleaning and reduce the omission.Based on the radar images,hierarchical visualization of processed multi-level information fusion is realized.The experiments show that the method can clean the redundant data accurately and achieve the efficient fusion of multi-level information of big data in the digital image.

Scale‐wise interaction fusion and knowledge distillation network for aerial scene recognition

Aerial scene recognition(ASR)has attracted great attention due to its increasingly essential applications.Most of the ASR methods adopt the multi‐scale architecture because both global and local features play great roles in ASR.However,the existing multi‐scale methods neglect the effective interactions among different scales and various spatial locations when fusing global and local features,leading to a limited ability to deal with challenges of large‐scale variation and complex background in aerial scene images.In addition,existing methods may suffer from poor generalisations due to millions of to‐belearnt parameters and inconsistent predictions between global and local features.To tackle these problems,this study proposes a scale‐wise interaction fusion and knowledge distillation(SIF‐KD)network for learning robust and discriminative features with scaleinvariance and background‐independent information.The main highlights of this study include two aspects.On the one hand,a global‐local features collaborative learning scheme is devised for extracting scale‐invariance features so as to tackle the large‐scale variation problem in aerial scene images.Specifically,a plug‐and‐play multi‐scale context attention fusion module is proposed for collaboratively fusing the context information between global and local features.On the other hand,a scale‐wise knowledge distillation scheme is proposed to produce more consistent predictions by distilling the predictive distribution between different scales during training.Comprehensive experimental results show the proposed SIF‐KD network achieves the best overall accuracy with 99.68%,98.74%and 95.47%on the UCM,AID and NWPU‐RESISC45 datasets,respectively,compared with state of the arts.

Performance measure for image fusion considering region information

An objective performance measure for image fusion considering region information is proposed. The measure not only reflects how much the pixel level information that fused image takes from the source image, but also considers the region information between source images and fused image. The measure is meaningful and explicit. Several simulations were conducted to show that it accords well with the subjective evaluations.

A NOVEL ALGORITHM OF MULTI-SENSOR IMAGE FUSION BASED ON WAVELET PACKET TRANSFORM

In order to enhance the image information from multi-sensor and to improve the abilities of the information analysis and the feature extraction, this letter proposed a new fusion approach in pixel level by means of the Wavelet Packet Transform (WPT). The WPT is able to decompose an image into low frequency band and high frequency band in higher scale. It offers a more precise method for image analysis than Wavelet Transform (WT). Firstly, the proposed approach employs HIS (Hue, Intensity, Saturation) transform to obtain the intensity component of CBERS (China-Brazil Earth Resource Satellite) multi-spectral image. Then WPT transform is employed to decompose the intensity component and SPOT (Systeme Pour I'Observation de la Therre ) image into low frequency band and high frequency band in three levels. Next, two high frequency coefficients and low frequency coefficients of the images are combined by linear weighting strategies. Finally, the fused image is obtained with inverse WPT and inverse HIS. The results show the new approach can fuse details of input image successfully, and thereby can obtain a more satisfactory result than that of HM (Histogram Matched)-based fusion algorithm and WT-based fusion approach.

Production of High-Resolution Remote Sensing Images for Navigation Information Infrastructures

This paper introduces the image fusion approach of multi-resolutionanalysis-based intensity modulation (MRAIM) to produce the high-resolution multi-spectral imagesfrom high-resolution panchromatic image and low-resolution multi-spectral images for navigationinformation infrastructure. The mathematical model of image fusion is derived according to theprinciple of remote sensing image formation. It shows that the pixel values of a high-resolutionmulti-spectral images are determined by the pixel values of the approximation of a high-resolutionpanchromatic image at the resolution level of low-resolution multi-spectral images, and in the pixelvalae computation the M-band wavelet theory and the a trous algorithm are then used. In order toevaluate the MRAIM approach, an experiment has been carried out on the basis of the IKONOS 1 mpanchromatic image and 4 m multi-spectral images. The result demonstrates that MRAIM image fusionapproach gives promising fusion results and it can be used to produce the high-resolution remotesensing images required for navigation information infrastructures.

AUTOMATIC SEGMENTATION OF HIPPOCAMPAL SUBFIELDS BASED ON MULTI-ATLAS IMAGE SEGMENTATION TECHNIQUES

The volume of hippocampal subfields is closely related with early diagnosis of Alzheimer's disease.Due to the anatomical complexity of hippocampal subfields,automatic segmentation merely on the content of MR images is extremely difficult.We presented a method which combines multi-atlas image segmentation with extreme learning machine based bias detection and correction technique to achieve a fully automatic segmentation of hippocampal subfields.Symmetric diffeomorphic registration driven by symmetric mutual information energy was implemented in atlas registration,which allows multi-modal image registration and accelerates execution time.An exponential function based label fusion strategy was proposed for the normalized similarity measure case in segmentation combination,which yields better combination accuracy.The test results show that this method is effective,especially for the larger subfields with an overlap of more than 80%,which is competitive with the current methods and is of potential clinical significance.

Information Fusion Methods in Computer Pan-vision System

Aiming at concrete tasks of information fusion in computer pan vision (CPV) system, information fusion methods are studied thoroughly. Some research progresses are presented. Recognizing of vision testing object is realized by fusing vision information and non vision auxiliary information, which contain recognition of material defects, intelligent robot’s autonomous recognition for parts and computer to defect image understanding and recognition automatically.

A Distributed Compressed Sensing for Images Based on Block Measurements Data Fusion

Compressed sensing (CS) is a new technique for simultaneous data sampling and compression. In this paper, we propose a novel method called distributed compressed sensing for image using block measurements data fusion. Firstly, original image is divided into small blocks and each block is sampled independently using the same measurement operator, to obtain the smaller encoded sparser coefficients and stored measurements matrix and its vectors.? Secondly, original image is reconstructed using the block measurements fusion and recovery transform. Finally, several numerical experiments demonstrate that our method has a much lower data storage and calculation cost as well as high quality of reconstruction when compared with other existing schemes. We believe it is of great practical potentials in the network communication as well as pattern recognition domain.

Lumbar spine localisation method based on feature fusion

To eliminate unnecessary background information,such as soft tissues in original CT images and the adverse impact of the similarity of adjacent spines on lumbar image segmentation and surgical path planning,a two‐stage approach for localising lumbar segments is proposed.First,based on the multi‐scale feature fusion technology,a non‐linear regression method is used to achieve accurate localisation of the overall spatial region of the lumbar spine,effectively eliminating useless background information,such as soft tissues.In the second stage,we directly realised the precise positioning of each segment in the lumbar spine space region based on the non‐linear regression method,thus effectively eliminating the interference caused by the adjacent spine.The 3D Intersection over Union(3D_IOU)is used as the main evaluation indicator for the positioning accuracy.On an open dataset,3D_IOU values of 0.8339�0.0990 and 0.8559�0.0332 in the first and second stages,respectively is achieved.In addition,the average time required for the proposed method in the two stages is 0.3274 and 0.2105 s respectively.Therefore,the proposed method performs very well in terms of both pre-cision and speed and can effectively improve the accuracy of lumbar image segmentation and the effect of surgical path planning.

基于信息增强和掩码损失的红外与可见光图像融合方法

针对低光场景下红外与可见光融合图像中存在的细节弱化和边缘模糊等问题,提出一种基于信息增强和掩码损失的图像融合方法。首先,采用引导滤波增强可见光图像的纹理细节和红外图像的边缘梯度;其次,构建双分支特征提取网络提取不同模态图像的特征信息,并设计交互增强模块以渐进交互的方式集成不同特征分支的互补信息,增强特征的细节表示;然后,在融合阶段设计注意力引导模块从空间和通道维度上关注特征信息,提升网络对关键特征的感知能力;最后,提出一种掩码损失以指导融合网络有针对性地保留源图像信息,提升融合质量。为验证所提方法的融合性能,在MSRS、TNO和LLVIP公开数据集上与9种主流的融合算法进行实验对比。结果表明,所提方法在定性和定量评估上均优于其它对比算法,生成的融合图像具有丰富的纹理细节、清晰的显著性目标和良好的视觉感知。

拉普拉斯卷积的双路径特征融合遥感图像智能解译方法

由于遥感图像存在多尺度变化和目标边缘模糊等问题,对其进行智能解译仍然是一项极具挑战性的工作。传统的语义分割方法在处理这些问题时存在局限性,难以有效捕捉全局和局部信息。针对上述问题,文中提出一种双路径特征融合分割方法 DFNet。首先,使用Swin Transformer作为主干提取全局语义特征,以处理像素之间的长距离依赖关系,从而促进对图像中不同区域相关性的理解;其次,将拉普拉斯卷积嵌入到空间分支,以捕获局部细节信息,加强目标地物边缘信息表达;最后,引入多尺度双向特征融合模块,充分利用图像中的全局和局部信息,以增强多尺度信息的获取能力。在实验中,使用了三个公开的高分辨率遥感图像数据集进行验证,并通过消融实验验证了所提模型不同模块的作用。实验结果表明,所提方法在Uavid数据集、Potsdam数据集、LoveDA数据集的mIoU达到了71.32%、85.58%、54.01%,提高了语义分割的性能,使分割结果更为精细。

基于结构功能交叉神经网络的多模态医学图像融合

针对多模态医学图像融合中存在纹理细节模糊和对比度低的问题,提出了一种结构功能交叉神经网络的多模态医学图像融合方法。首先,根据医学图像的结构信息和功能信息设计了结构功能交叉神经网络模型,不仅有效地提取解剖学和功能学医学图像的结构信息和功能信息,而且能够实现这两种信息之间的交互,从而很好地提取医学图像的纹理细节信息。其次,利用交叉网络通道和空间特征变化构造了一种新的注意力机制,通过不断调整结构信息和功能信息权重来融合图像,提高了融合图像的对比度和轮廓信息。最后,设计了一个从融合图像到源图像的分解过程,由于分解图像的质量直接取决于融合结果,因此分解过程可以使融合图像包含更多的细节信息。通过与近年来提出的7种高水平方法相比,本文方法的AG,EN,SF,MI,QAB/F和CC客观评价指标分别平均提高了22.87%,19.64%,23.02%,12.70%,6.79%,30.35%,说明本文方法能够获得纹理细节更清晰、对比度更好的融合结果,在主观视觉和客观指标上都优于其他对比算法。

MCFNet:融合上下文信息的多尺度视网膜动静脉分类网络

针对由于血管类间具有强相似性造成的动静脉错误分类问题,提出了一种新的融合上下文信息的多尺度视网膜动静脉分类网络(multi-scale retinal artery and vein classification network,MCFNet),该网络使用多尺度特征(multi-scale feature,MSF)提取模块及高效的全局上下文信息融合(efficient global contextual information aggregation,EGCA)模块结合U型分割网络进行动静脉分类,抑制了倾向于背景的特征并增强了血管的边缘、交点和末端特征,解决了段内动静脉错误分类问题。此外,在U型网络的解码器部分加入3层深度监督,使浅层信息得到充分训练,避免梯度消失,优化训练过程。在2个公开的眼底图像数据集(DRIVE-AV,LES-AV)上,与3种现有网络进行方法对比,该模型的F1评分分别提高了2.86、1.92、0.81个百分点,灵敏度分别提高了4.27、2.43、1.21个百分点,结果表明所提出的模型能够很好地解决动静脉分类错误的问题。

双波段彩色融合图像色彩和谐性客观评价

针对现有的图像质量评价方法较少利用人眼视网膜和视觉皮层的颜色编码机制,并且未能充分考虑图像色彩信息对图像质量的影响,提出了一种基于多视觉特征的可见光(微光)与红外彩色融合图像色彩和谐性客观评价模型.该模型在图像质量评估中融入了更多的颜色信息,综合考虑多种人眼视觉特征包括视觉对立色彩特征、色彩信息波动特征和高级视觉内容特征,经过特征融合和支持向量回归训练,实现彩色融合图像的色彩和谐性客观评价.采用3种典型场景融合图像数据库进行实验比较与分析.实验结果表明,与现有的8种图像质量客观评价方法相比,所提出的方法与人眼主观感受更加一致,具有较高的预测准确度.

基于多孔卷积神经网络的图像空间结构信息细节表征

针对传统图像空间结构信息表征方法存在细节表征模糊度较高、信息训练损失较高等问题,提出一种新的基于多孔卷积神经网络的图像空间结构信息细节表征方法。该方法通过图像空间结构信息细节相似性度量,并以图像的形状、颜色和纹理特征对图像空间结构信息细节进行编码,再去除图像冗余信息,利用多孔卷积神经网络对图像空间结构的深度信息进行融合,从而完成图像空间结构信息的细节表征。实验结果表明,基于多孔卷积神经网络的图像空间结构信息细节表征方法在模糊度、训练损失、图像相似性等方面都比传统的3种方法优越,能够清晰地表征图像空间结构信息。

基于文本-图像增强的突发事件识别及分类方法研究

[目的/意义]丰富的互联网数据为洞悉真实事件提供了多维视角,快速识别突发事件并准确判断其所属类别,有助于各级政府及应急管理部门高效地管理应急情报资源。[方法/过程]文章构建了基于文本—图像增强的突发事件识别及分类的理论模型;通过文本卷积神经网络、视觉几何群网络搭建深度神经网络共同组成Multi-DNN模型;最后以真实的自然灾害类突发事件数据进行实例验证。[结果/结论]通过文本、图像相互增强,多模态特征融合能够提升突发事件识别及分类的准确率,同时在小样本数据的任务处理中仍有良好效果,证明不同模态的数据能够相互补充、相互印证,对其融合处理能够提供比单一模态更为准确和全面的信息分析。

一种大孔径静态干涉高光谱成像数据压缩方法

大孔径静态干涉成像遥感数据的数据量较大,需要寻找一种合适的方法对其压缩。从大孔径静态干涉成像机理出发,分析了干涉数据的空间和干涉维冗余性,并基于现有成熟混合压缩编码方法,提出了基于相似干涉曲线与不同光程差之间冗余去除的算法,对冗余数据进行去除预处理。对干涉数据进行基于曲线表的干涉曲线编码表示,对不同光程差之间的图像进行相关性预测,减少了大孔径静态干涉成像遥感图像的量化深度并降低了图像的信息熵,再结合JPEG2000算法进行无损或有损压缩。实验结果表明,对于大孔径静态干涉成像数据,该算法可实现压缩比为3.1倍的无损压缩,有损压缩的率失真曲线也优于其他对比算法,其复原图像反演出的光谱曲线的光谱角和相对二次误差均优于其他对比算法处理的数据,有效保护了光谱信息。

基于Landsat8卫星光谱与纹理信息的森林蓄积量估算

以福建省将乐县国有林场为研究对象,通过外业实地调查得到样地蓄积量:以Landsat 8卫星遥感图像为数据源,对遥感图像进行处理,获取多光谱影像的波段光谱值、植被指数和波段组合值,并筛选出全色波段的最优纹理生成窗口与纹理特征;通过多元回归分析方法,分别建立仅以光谱因子为自变量和结合光谱信息和纹理特征的蓄积量估测模型,并比较两者之间的精度。实验结果表明:光谱因子的多元线性回归方程的相关系数为0.853,联合光谱和纹理特征因子反演的多元回归方程的相关系数为0.926。同时利用检验数据,得出模型的预测精度:光谱因子蓄积量的估算方程精度为79.81%,联合反演蓄积量的估算方程精度为85.98%。研究表明:引入纹理特征后蓄积量的预测精度得到一定程度的提高,利用Landsat 8全色波段的纹理特征进行蓄积量估测具有良好的应用前景。

遥感图像不同融合方法的适应性评价——以ZY-3和Landsat8图像为例

目前对常用融合方法用于ZY-3和Landsat8卫星图像的适用性研究较少,为此,探讨了小波变换法(WT法)、相位恢复法(Gram_Schimdt,G-S法)、彩色标准化变换法(Brovey法)、主成分变换法(PCA法)、彩色空间变换法(IHS法)以及超分辨率贝叶斯法(Pansharp法)等6种融合方法对ZY-3和Landsat8图像的适用性,并从"光谱信息保真度"和"空间信息融入度"2个方面对融合图像进行了评价。结果表明:在空间信息融入方面,对于ZY-3图像,Pansharp法的信息融入度最差,IHS法的信息融入度最好,PCA法、Brovey法、G-S法和WT法次之;对于Landsat8图像,G-S法的信息融入度最好,WT法最差。在光谱信息保真方面,对于ZY-3图像,PCA法具有较高的光谱保真度,IHS法、G-S法和Brovey法次之,WT法最差;对于Landsat8图像,G-S法具有较高的光谱保真度,Pansharp法和Brovey法次之,IHS法、WT法和PCA法较差。