This paper presents an approach for attributed scattering center (ASC) matching with application to synthetic aperture radar (SAR) automatic target recognition (ATR). A statistics-based distance measure is designed evaluate the between individual ASCs. Afterwards, Hungarian algorithm employed build a one-to-one correspondence two ASC sets. Based on correspondence, global similarity and local are comprehensively consistency structural correlation those The similarities exploit inner sets,...

Multimodal image registration is the fundamental technique for scene analysis with series remote sensing images of different spectrum region. Due to highly nonlinear radiometric relationship, it quite challenging find common features between modal types. This paper resorts deep neural network, and tries learn descriptors multimodal patch matching, which key issue registration. A Siamese fully convolutional network set up trained a novel loss function, adopts strategy maximizing feature...

The quality of synthetic aperture radar (SAR) images and the completeness template database are two important factors in template-based SAR automatic target recognition. This letter gives a solution to by multilevel reconstruction targets using attributed scattering centers (ASCs). ASCs original extracted reconstruct target's image, which not only reduces noise background clutters but also keeps electromagnetic characteristics target. Template reconstructed at multilevels simulate various...

Feature extraction and matching are two important steps in synthetic aperture radar automatic target recognition. This paper uses the binary region as feature proposes a scheme for regions using morphological operations. The residuals between testing its corresponding template processed by opening operation. Then, similarity measure is defined based on residual remains to evaluate similarities different targets. Afterward, Bayesian decision fusion employed fuse gained structuring elements...

Automatic target recognition (ATR) of synthetic aperture radar (SAR) images is performed on either global or local features. The features can be extracted and classified with high efficiency. However, they lack the reasoning capability thus hardly work well under extended operation conditions (EOCs). are often more difficult to extract classify but provide for EOC recognition. To combine efficiency robustness in an ATR system, a hierarchical fusion proposed SAR this paper. As features,...

This paper proposes a robust method for the matching of attributed scattering centers (ASCs) with application to synthetic aperture radar automatic target recognition (ATR). For testing image be classified, ASCs are extracted match ones predicted by templates. First, Hungarian algorithm is employed those two ASC sets initially. Then, precise carried out through threshold method. Point similarity and structure calculated, which fused evaluate overall based on Dempster–Shafer theory evidence....

The coregistration of optical and synthetic aperture radar (SAR) imageries is the bottleneck in exploring complementary information from two multimodal datasets. difficulties lie not only complex radiometric relationship between them, but also distinct geometrical models SAR imaging systems, which cause it nontrivial to explicitly depict spatial corresponding image regions when elevation fluctuations exist. This article aims investigate flow technique for pixelwise dense registration...

Azimuth sensitivity is a significant characteristic of synthetic aperture radar (SAR) images. Most the previous SAR target recognition algorithms try to cope with property by pose estimation or training classifiers which are not sensitive azimuth. Actually, azimuth can provide discriminative information for as supplement original spatial image (SI). This letter describes (ASI) constructed comparing sub-aperture images SI. Then SI and ASI classified sparse representation-based classification...

Multi‐feature decision fusion is a robust way to enhance the performance of synthetic aperture radar (SAR) automatic target recognition (ATR). This study fuses two physically relevant features namely attributed scattering centres and contour for SAR ATR. The provide an intuitive comprehensive description targets which can not only promote ATR but also benefit human interpretation. complement each other with weak spatial correlation. Furthermore, both be concisely represented by simple point...

Post-processing methods that rely on fusion-grided forecast products can reduce systematic biases from Numerical Weather Prediction (NWP) precipitation forecasts. However, these also limit the capability to accurately at local stations. We constructed a Station-based Precipitation Model (SPPM) utilizes deep-learning algorithms, predominantly convolutional layers and ResNet modules. Based 390 meteorological stations in North China European Centre for Medium-Range Forecasts Highest-resolution...

Vehicle re-identification (ReID) aims to identify a target vehicle in different cameras with non-overlapping views, and it plays an important role when the car licence plate recognition is unavailable or unreliable. Compared face person ReID tasks, difficult train effective model due two reasons: views greatly affect visual appearance of vehicle, vehicles may exhibit fairly similar their images are captured from one unified single view. To handle these training difficulties, we introduce...

In wideband stepped-frequency radar systems, relative motions between and target can induce high distortions of high-resolution range profiles (HRRPs), such as migration, shape deformation, signal-to-noise-ratio (SNR) loss. These distortions, if ignored, lead to unacceptable performance deterioration in detection. To solve this problem, a new algorithm for detecting moving range-spread targets (RSTs) is proposed letter. The detector utilizes the waveform contrast HRRP perform both motion...

A three-dimensional electromagnetic model (3-D EM-model)–based scattering center matching method is developed for synthetic aperture radar automatic target recognition (ATR). 3-D EM-model provides a concise and physically relevant description of the target’s phenomenon through its centers which makes it an ideal candidate ATR. In our method, scatters are projected to two-dimensional measurement plane predict scatters’ location intensity properties. Then identical information extracted in...

1ReH•Cl, a highly robust and antiaromatic rhenium(I) complex of triarylrosarin, is synthesized. The 1H NMR spectrum 1ReH•Cl shows upfield-shifted pyrrole protons downfield-shifted inner that confirm its nature, with density functional theory calculations strongly supporting this interpretation. Antiaromatic absorbs from the UV to near-IR region optical spectrum; cyclic voltammetry, thin-layer UV-vis spectroelectrochemistry, spin-density distributions clearly reveal rosarin backbone undergoes...

This paper proposes a hierarchical control strategy that achieves the power management and coordinated of grid-connected DC microgrids. The proposed architecture consists two levels. In primary level, State-of-charge (SOC) based droop is applied to interfacing converters storage elements (SE). slope adjusted according individual SOC, aiming dynamic sharing SOC balance among multiple SE. Vdc-P instead P-f implemented in interlinking converter between subgrid main grid, thus bidirectional...

Due to the extremely complex composition of remote sensing scenes, REmote Sensing Image Scene Classification (RE-SISC) is still a challenging task.To further improve classification accuracy, this article introduces deep-learning detector into RE-SISC and proposes classify images according detected class-specific signature objects.Inspired by procedure human vision system, we design framework that utilizes objects scene classes guide classification.When performing image classification,...

The pose of a rotational symmetry target in wideband radar is the angle between its axis and line sight. It required parameter for absolute attitude measurement using bistatic measuring system. In this letter, 3-D electromagnetic model (3-D EM model)-based method proposed estimating from fully polarimetric measurements. established offline based on geometric structure. Scattering features at different poses can be accurately predicted by model. algorithm, synthetic high-resolution range...

Road extraction from high-resolution remote sensing images has been a challenging problem in ground object detection field. In this paper, an effective and anti-noise road method by using wavelet transform Hough is proposed. Firstly, used for image denoising detecting edges of roads. Then the segmented into area non-road area. to extract roads after segmentation. Finally, effectiveness testified results experiment.

Vehicle re-identification (re-ID) is to identify the same vehicle captured by different cameras with non-overlapping views, which plays an important role in intelligent transportation system and traffic safety. Compared face recognition person re-ID tasks, it difficult train effective model since vehicles of model, such as Mercedes-Benz C300, may exhibit strong inter-class similarity. To handle this difficulty, we propose a multi-grain ranking loss auxiliary models task two sub-tasks...

This paper proposes a new feature extraction method for synthetic aperture radar (SAR) images with application to automatic target recognition (ATR). The original SAR image is first represented by sparse containing only few dominant scattering centers (SCs). According the theory of 2D compressive sensing (CS), can be reconstructed from low dimensional matrix little distortion. Therefore, we use random projection extract features image. proposed directly works on thus avoiding conversion...

The similarity evaluation between three‐dimensional (3D) electromagnetic model (em‐model) and 2D synthetic aperture radar (SAR) data is a key factor in em‐model‐based SAR automatic target recognition (ATR). In this study, scatterer‐based approach proposed to evaluate the 3D em‐model for ATR purpose. A characterised by set of scatterers evaluated through all these scatterers. First, information each scatterer predicted from used guide extraction data. Then pair hypothesis testing approach....

In this paper, we propose a fast mosaicking method for an image sequence acquired by small UAV. It can be divided into three steps. First, homographies between images are estimated. Then, based on the estimated homographies, bundle adjustment using number of GCPs is applied calculating orthorectification parameters. Finally, all UAV rectified according to calculated parameters and merged together establish mosaic. Experiments data measured real carried out validate effectiveness proposed method.