Search

Gesture Recognition which is an inevitable part of human computer interaction is an ever evolving active research area. This has been in use under varied applications like smart home systems, sign language recognition, augment reality and in device controls. The acquisition of hand gestures are usually through optical sensors, in this work a radar based hand gesture data set is used for classifying the gestures. Micro doppler signatures were used as the input to the model. Radar dataset has fewer data samples compared to optical based data sets. The proposed work uses separable convolutional neural networks model which does depth wise convolution followed by point wise convolution to reduce overfitting effect of the training data. The proposed model was built in such a way that the model is capable of classifying any unseen data without exactly mimicking the training samples. The proposed model has achieved 94.56% as the testing accuracy which is certainly better than the previous work on this Dop Net data set. Moreover the model has also minimized the computational hours of the model using separable convolutions.

Excavation is common in the construction of hydropower projects. Monitoring for both stable and unstable slopes is a pressing requirement during progressive excavations with or without cut-slope support measures. The present study is about an excavation carried out in a proposed concrete gravity dam site on the right bank slope of the Punatsangchhu River in western Bhutan. During the excavation, the right bank abutment witnessed multiple events of slope failures of various magnitudes. One major landslide occurred on 23 July 2013 in the toe of the right abutment, where foliation/multiple sheared/fractured zones/seams in quartz-feldspathic biotite gneiss dip towards the valley. During further excavations/piling works, a subsequent landslide on the downstream, within the body of the July 2013 landslide, occurred on 12 August 2016. As a result, a real-time monitoring of the slope became necessary to facilitate further excavations for achieving the dam foundation. Thus, the advanced slope monitoring instrument like Image By Interferometric Survey – Frequency Modulated (IBIS-FM) radar was deployed for monitoring the right bank slope during the excavation for the construction of ∼129 m high dam. The displacement of the hill slope is assessed based on this system by monitoring point locations as well as areas by assigning different threshold values for providing timely alerts. This real-time monitoring was effective in identifying the reactivation of August 2016 landslide that occurred on 22 January 2019. Thus, this study showcases the efficiency of IBIS-FM radar in monitoring slope instability with sub-millimeter accuracy on a near real-time basis.

In low signal-to-noise ratio (SNR) environments, the traditional radar emitter recognition (RER) method struggles to recognize multiple radar emitter signals in parallel. This paper proposes a multi-label classification and recognition method for multiple radar-emitter modulation types based on a residual network. This method can quickly perform parallel classification and recognition of multi-modulation radar time-domain aliasing signals under low SNRs. First, we perform time-frequency analysis on the received signal to extract the normalized time-frequency image through the short-time Fourier transform (STFT). The time-frequency distribution image is then denoised using a deep normalized convolutional neural network (DNCNN). Secondly, the multi-label classification and recognition model for multi-modulation radar emitter time-domain aliasing signals is established, and learning the characteristics of radar signal time-frequency distribution image dataset to achieve the purpose of training model. Finally, time-frequency image is recognized and classified through the model, thus completing the automatic classification and recognition of the time-domain aliasing signal. Simulation results show that the proposed method can classify and recognize radar emitter signals of different modulation types in parallel under low SNRs.

As small satellites are finding increased applicability in a number of space missions, the community is actively exploring the feasibility of missions employing active instruments on small spacecraft. In particular, synthetic aperture radars are of interest as they prove to be versatile tools for Earth Observation purposes. This paper proposes a trade space exploration model to identify designs for space-borne satellite synthetic aperture radars that address the desired instrument requirements. Two trade space analysis approaches are proposed, based on a multivariate analysis with two evaluation metrics and using parallel coordinate plots. The paper demonstrates the utility of the trade space model by considering two case studies. The first study looks at radar instruments on a broad range of small satellite platforms. This study narrows down a trade space of 1265 feasible radar designs to less than 44 Pareto optimal designs at different center frequencies (C- and X-band), elicits conditions for L-band SARs feasibility, and discusses the feasibility bounds and technical constraints on associated instrument and spacecraft requirements. The second study considers radars for 3U CubeSat platforms. The study identifies 44 Pareto optimal designs within 12,928 feasible designs and sheds light on the operational constraints required for the development of such innovative miniaturized radars.

This paper considers the problem of target and jamming recognition for the pulse Doppler radar fuze (PDRF). To solve the problem, the matched filter outputs of the PDRF under the action of target and jamming are analyzed. Then, the frequency entropy and peak-to-peak ratio are extracted from the matched filter output of the PDRF, and the time-frequency joint feature is constructed. Based on the time-frequency joint feature, the naive Bayesian classifier (NBC) with minimal risk is established for target and jamming recognition. To improve the adaptability of the proposed method in complex environments, an online update process that adaptively modifies the classifier in the duration of the work of the PDRF is proposed. The experiments show that the PDRF can maintain high recognition accuracy when the signal-to-noise ratio (SNR) decreases and the jamming-to-signal ratio (JSR) increases. Moreover, the applicable analysis shows that he ONBCMR method has low computational complexity and can fully meet the real-time requirements of PDRF.

Traditional clustering algorithms can be applied for the pre-sorting step of radar signal sorting. It can effectively dilute the pulse stream and prevent the dense pulse stream from interfering pulse repetition interval (PRI) extraction. However, the pre-sorting deviation will cause interference and missing pulses during the main sorting process. To solve this problem, we deploy the unmanned aerial vehicle (UAV) swarm to monitor reconnaissance areas and put forward a novel deep transfer learning based signal sorting method. The UAV swarm can collect the pulses from different time and spatial domains, and interference and missing pulses in main sorting processing can be relieved dramatically. In our model, we pre-train our model with the data collected from multiple source areas, which corresponds to different areas detected by different parts of UAV swarms. Then we fine-tune our model with the data of the target area. The experimental results prove that the signal sorting accuracy of methods based on deep transfer learning, i.e., YOLO-MobileNet, F-RCNN and cascade RCNN, are higher than that of the baseline methods. In addition, the signal sorting accuracy of traditional methods based on deep learning can be greatly improved with the help of transfer learning.

Synthetic aperture radar (SAR) is playing an important role in maritime domain awareness. As a fundamental ocean mission, SAR ship detection can offer high-quality services for marine surveillance. It receives widespread concern in the SAR remote sensing community. Recently, deep learning (DL) has greatly improved SAR ship detection accuracy. However, there are still some unperceived imbalance problems that seriously hinder further accuracy improvements among current DL-based SAR ship detectors. Therefore, we propose a novel concept of balance learning (BL) for high-quality SAR ship detection. We first point out the four unperceived imbalance problems, i.e., image sample scene imbalance, positive negative sample imbalance, ship scale feature imbalance, and classification regression task imbalance. Then, we offer some profound insights into these imbalances. Immediately, we propose four effective solutions to handle the above four imbalances correspondingly, i.e., balance scene learning mechanism (BSLM), balance interval sampling mechanism (BISM), balance feature pyramid network (BFPN), and balance classification regression network (BCRN). Finally, combined with the four solutions, a novel balance learning network (BL-Net) is proposed. Ablation studies can confirm each solution’s effectiveness. Experimental results on five open datasets (SSDD, Gaofen-SSDD, Sentinel-SSDD, SAR-Ship-Dataset, and HRSID) reveal BL-Net’s state-of-the-art SAR ship detection performance compared to the other 30 DL-based SAR ship detectors. Specifically, in contrast to the current most competitive method, the accuracy increase of BL-Net is 2.98% on SSDD, 1.97% on Gaofen-SSDD, 1.49% on Sentinel-SSDD, 0.55% on SAR-Ship-Dataset, and 4.95% on HRSID. Last but not least, the satisfactory ship detection results on another two large-scene Sentinal-1 SAR images confirm BL-Net’s strong migration capability. This indicates BL-Net’s potential value in marine surveillance.

Synthetic aperture radar (SAR) image recognition is an important stage of SAR image interpretation. The standard convolutional neural network (CNN) has been successfully applied in the SAR image recognition due to its powerful feature extraction capability. Nevertheless, the CNN requires numerous labeled samples for satisfactory recognition performance, while the performance of the CNN decreases greatly with insufficient labeled samples. Aiming at improving the SAR image recognition accuracy with a small number of labeled samples, a new few-shot learning method is proposed in this paper. We first utilize the attention prototypical network (APN) to calculate the average features of the support images from each category, which are adopted as the prototypes. Afterwards, the feature extraction is performed on the query images using the attention convolutional neural network (ACNN). Finally, the feature matching classifier (FMC) is adopted for calculating the similarity scores between the feature maps and the prototypes. We embed the attention model SENet to the APN, ACNN, and FMC, which effectively enhances the expression of the prototypes and the feature maps. Besides, the loss function of our method consists of cross-entropy and prototype-separability losses. In the training process, this loss function increases the separability of different prototypes, which contributes to higher recognition accuracy. We perform experiments on the Moving and Stationary Target Acquisition and Recognition (MSTAR) and the Vehicle and Aircraft (VA) datasets. It has been proved that our method is superior to the related state-of-the-art few-shot image recognition methods.

Precise localization and tracking of moving non-collaborative persons and objects using a network of ultra-wideband (UWB) radar nodes has been shown to represent a practical and effective approach. In UWB radar sensor networks (RSNs), existence of strong clutter, weak target echoes, and closely spaced targets are obstacles to achieving a satisfactory tracking performance. Using a track-before-detect (TBD) approach, the waveform obtained by each node during a time period are jointly processed. Both spatial information and temporal relationship between measurements are exploited in generating all possible candidate trajectories and only the best trajectories are selected as the outcome. The effectiveness of the developed TBD technique for UWB RSNs is confirmed by numerical simulations and by two experimental results, both carried out with actual UWB signals. In the first experiment, a human target is tracked by a monostatic radar network with an average localization error of 41.9 cm with no false alarm trajectory in a cluttered outdoor environment. In the second experiment, two targets are detected by multistatic radar network with localization errors of 25.4 cm and 19.7 cm, and detection rate of the two targets is 88.75%, and no false alarm trajectory., 39 pages, 20 figures. Accepted in Signal Processing (Elsevier)

The detection and positioning of unmanned aerial vehicles has become essential for both automation and surveillance tasks, in recent years. The design of accurate drone localization systems is challenging, especially in cluttered environments, where the target may be partially or even completely obscured. This paper proposes a precise detection and 3D localization system for drones, by means of a millimetre wave radar. Drone locations are estimated from spatial heatmaps of the received radar signals, which are obtained by applying the super-resolution MUSIC algorithm. These estimates are improved by analysis of the micro-Doppler effect, generated by the rotating propellers, which aids detection in poor visibility conditions. A novel Gaussian Process Regression model is developed, in order to compensate for systematic biases in the radar data. The complete system produces accurate estimates of the target range and direction, and is shown to outperform direct spectral analysis methods.

The automotive industry is interested in the estimation of vehicle trailer rotation (trailer angle or hitch angle) due to its use in trailer control algorithms. We present an experimental dataset collected in a study of the estimation problem [1] and a MATLAB code implementation of the study. The data collection apparatus is a truck mock-up that is attached to a flatbed trailer at the hitch ball. Two radars are installed in the taillight fixtures of the truck and a camera is installed in the truck’s tailgate like a typical backup camera installation. A rotary motion sensor is also installed at the hitch ball to provide ground truth measurement of the trailer angle. To aid analysis of the dataset, both radar detections are transformed onto a vehicle coordinate system (VCS) having its origin at the hitch ball i.e. the different radar viewpoints are combined into one with respect to the hitch ball. The MATLAB code presented with this article has two major functionalities. The first functionality is the visualization of both radar detections, the combined radar detections in the VCS, the camera images, and the ground truth angles, as the trailer rotates. The second functionality of the code is the replication of the estimation results in [1], which used only the radar detections from the dataset.

We fabricated a ZrO2/Al2O3 ceramic using the 3D plasma spraying technology, and the surface morphology and electrical properties of the prepared ceramic were investigated. The microwave-absorbing performance of the ceramic was studied based on the measured electrical properties. Furthermore, a radar-absorbing structure (RAS) was designed based on the fabricated ZrO2/Al2O3 ceramic with a K0.5Na0.5NbO3 (KNN) film heterojunction. The relationship between the structural parameters of the RAS and the microwave performance was explored. The changes in the absorption performance and band structure of the RAS with the resistance of the KNN film were investigated. The analytical results showed that the structural parameters of the KNN film can facilitate the absorption intensity as well as the bandwidth of the RAS while inducing electromagnetic coupling in the RAS system. The improvement in the absorption performance could be attributed to the resonance coupling between the ZrO2/Al2O3 ceramic and the KNN film.

Coastal grasslands are semi-natural habitats used for agriculture, but they are also vital habitats for many taxa, such as waders. The importance of this habitat for wildlife has found recognition in the policies of the European Union, resulting in the implementation of biodiversity-friendly agri-environment schemes (AES) to ensure the sustainable management of these areas. However, the performance of AESs has often been questioned and data to analyse their effectiveness is mostly lacking. To fill this knowledge gap, we used freely available European Space Agency's Sentinel-1 Synthetic Aperture Radar (SAR) images and ground truth data to describe the vegetation height based habitat suitability of coastal grasslands as breeding sites for waders at a country-wide scale in Estonia, where the main threat to this habitat is overgrowing with high sward plant species. We show that satellite remote sensing can be used as a straight-forward and reliable tool to estimate the grazed coastal grassland habitat suitability for breeding waders in terms of vegetation height in two broad classes. We demonstrate that the AES applied on Estonian coastal grasslands is important for improving the suitability of this habitat for the breeding wader community: compared to 24% of suitable habitat in areas where no agricultural subsidies are paid, 60% of the land under the targeted AES is suitably managed for breeding waders. The highest proportion of habitat suitability (76%) is achieved in areas receiving targeted top-up subsidy for wader conservation. Continued monitoring of the performance of the AES is essential for taking timely and targeted conservation actions and we strongly recommend the wider use of remotely sensed data for this purpose. We promote the continuation of the AES scheme applied on Estonian coastal grasslands while improving its efficiency through carefully raised stocking rates and increased uptake of the top-up scheme.

This study aims to solve the problem of material detection and shovel difficulty judgment of earth moving machinery and characterize the material shovel difficulty by using radar chart method comprehensively. YOLOv4 object detection network is used for material detection, and the average detection accuracy reaches 93.03%. We fuse deep learning technology and radar chart method information and explore the correlation between material detection and material shoveling ease assessment. The evaluation of the identifiable shovel difficulty of the material is achieved with vision sensors, and the method is real-time and efficient. This is an incipient attempt. The detailed information provided in this study is valuable for the autonomous shoveling of loaders.

Radar-based life detecting technology has been researched more and more in recent years and applied widely in several areas. However, few researches on multi-subjects detecting and locating have been reported. Even these researches are mainly based on imaging technology with two shortcomings, namely the interference of ghost and difficult to focus the target’s position. A new coarse-to-fine approach for detecting and locating of human targets using single-input multiple-output (SIMO) radar under penetration conditions is proposed. The approach realized automatic recognition and localization of human subjects without any prior information, such as target count. Firstly, an improved censored mean level detector constant false alarm rate (CMLD-CFAR) by adding a feature to distinguish vital sign from background is proposed in each inner-channel to coarsely determine the targets’ time-of-arrival (TOA) distances. Secondly, the real targets count is finely determined by calculating the inter-channel Pearson correlation coefficient of signals from different channels. Then, a back projection (BP) is presented to calculate the targets’ 2-D locations. In the experiments of single and multi-target locating, all targets were identified correctly and localized near their actual positions with errors within 0.3 m. With the distinctive telescopic structure of the proposed radar, the azimuth resolution of locating is adjustable. Compare to current radar system for radar multi-target locating, like multiple-input multiple-output (MIMO) system, the proposed system can improve accuracy and automation of localization while decline the cost and complexity. It has potential to promote practical application of locating trapped human targets in post-disaster rescue and intelligent homeware fields.

Deep-learning-based gait classification of young and elderly adults using micro-Doppler radar (MDR) is presented in this paper. The MDR signal data were accurately simulated using an open motion-capture gait dataset, and deep-learning classification of the time-velocity distribution (i.e., spectrogram) images calculated with the generated data are presented. Utilizing a simulation, we also investigated the body parts deemed most efficient for classification based on their generation of good MDR data. As a result, the classification rate using whole-body data was 74%. However, this classification rate of using only leg data showed an accuracy of 91%, which indicates that the thighs and shanks are efficient target body parts for the gait classification of both young and elderly adults.

Accurate crop mapping for agricultural monitoring requires the use of robust image classification algorithms. Cultivation of illegal cannabis is common in the Bekaa plain, Lebanon, as in some other areas of the world (Morocco, Afghanistan). Identification of cultivation sites is money and time consuming since it highly relies on knowledge of cultivation areas and field surveys. Machine-learning-based image classification provides an alternative method for cannabis detection. This paper presents a comparative analysis of four machine-learning classifiers implemented in Google Earth Engine: Random Forest (RF), Gradient Boosting (GBT), Classification and Regression Tree (CART), and Support Vector Machine (SVM) for cannabis and other crop type classification. We implement and test several image fusion approaches for optical (Sentinel-2, Landsat) and Synthetic Aperture Radar (SAR) imagery (Sentinel-1) along with several combinations of surface textural features from Sentinel-1. Six different crop groups were classified over three years (2016, 2017, and 2018) in the Bekaa valley of Lebanon. In general, although SVM outperformed RF, GBT, and CART classifiers with an overall classification accuracy of more than 90%, RF and GBT provided more consistent results in terms of the cannabis cultivated area. SVM appears to be sensitive to the size of the training data. RF and GBT estimate an average cannabis area of 2800 ha in 2016, 2983 ha in 2017, and 5900 ha in 2018. Our results demonstrate that the fusion of radar and optical imagery can improve image classification accuracy by up to 5%. A marginal improvement in overall accuracy was observed when textural features were added. This is the first study that used image fusion and ML to estimate illegal cannabis cultivation areas in Lebanon and help evaluate the contribution of cannabis to the local economy. The combination of machine learning algorithms and imagery fusion proved reliable for crop classification in general and cannabis in particular.

Tool wear monitoring during the cutting process is crucial for ensuring part quality and productivity. A data-driven monitoring approach based on radar map feature fusion is proposed for tool wear recognition and quantitative prediction, aiming at tracking the evolution of tool wear comprehensively. Specifically, the sensitive features from multi-source signals are fused by a radar map, and health indicators capable of characterizing the tool wear evolution are obtained. For the recognition of tool wear state and the quantitative prediction of tool wear values, the Adaboost Decision Tree (Adaboost-DT) ensemble learning model and stacked bi-directional long short-term memory (SBiLSTM) deep learning network are established, respectively. Experimental results demonstrated that the proposed approach could recognize the current wear state quickly and accurately whilst predicting wear values based on limited historical data available. Combining tool wear recognition and prediction results contributes to making a more flexible tool replacement decision in intelligent manufacturing processes.

Sharpening ground-penetrating radar (GPR) images remains a persistent challenge. Two approaches in particular—singular value decomposition (SVD) and sparse blind deconvolution (SBD)—have been shown to effectively denoise GPR images and resolve a reflectivity model, respectively. To our knowledge, these approaches have not been used to process GPR data from archaeological sites. This paper highlights the treatment of a sample GPR profile acquired over two graves in sandy soil. SVD can remove user-specified components of the GPR image such as the direct wave, horizontal banding, and ringing from surface features. These are removed via a low order eigenimage filter in which the decomposed GPR data is reconstructing without the eigenimages containing them. This type of filter significantly alters the frequency content and total amplitude, which may or may not negatively influence subsequent filtering techniques or data analysis. Thus, we use SBD to investigate the influence of eigenimage filtering. The eigenimage-filtered GPR image returns a more physically meaningful reflectivity model compared to SBD of the original GPR profile. We additionally test the common approach ‘denoising’ GPR data by also omitting higher-order eigenimages from the reconstructed GPR image. The resulting reflectivity model was less sparse indicating that higher-order eigenimages contain signal information critical to deconvolution. To test the effectiveness of other filtering steps with SVD filtering, we compare the migration of the original GPR profile against that of an eigenimage-filtered image. The migrated filtered image, particularly if higher-order eigenimages are preserved, provides more spatially constrained point reflectors regardless of the migration algorithms used compared to the migrated original profile.

Conventional high resolution range profile (HRRP) based radar automatic target recognition (RATR) methods are based on a complete template library assumption. However, such an assumption can barely hold in real world applications since it is difficult to observe target echoes from all aspects, especially for noncooperative targets. To alleviate the target-aspect missing problem, we develop a domain-aware meta network (DOAMN) for HRRP-based RATR. Specifically, the DOAMN first uses a domain-aware (DOM) module to distinguish whether HRRPs come from the seen or unseen target aspects, then a meta network (MNet) is employed to learn a generalized parameter initialization that is able to achieve fast adaptation among target aspects. To enable the proposed DOAMN to be trained in an end-to-end manner, we further present an effective iterative hybrid optimization method. Experiments on simulated HRRP dataset demonstrate the effectiveness and efficiency of the proposed model.

Within the context of masonry texture identification and geometrical reconstruction, Ground Penetrating Radar (GPR) could emerge as an effective survey method, due to its high-resolution, scalable and non-destructive approach. This study has methodologically assessed operational advantages and weaknesses of the methodology when implemented for a seismic assessment purpose. In particular, the scope was to examine acquisition and post-processing strategies that could reliably highlight the presence or absence of specific construction features on which to develop a quantitative indication of the quality of the masonry. For this purpose, a high frequency 3D GPR acquisition on a plastered masonry wall incorporating different brick geometries, has been carried out. The acquisition was made along two different survey directions and analysing unfocussed and migrated radar slices. Thanks to a favourable antenna pattern alignment and achieved resolution enhancement, the focussed GPR data collected along the vertical mortar joints orientation have proven to be able to properly reconstruct the wall texture, providing therefore detailed and crucial information on its geometrical appearance.

Digital Pulse-Doppler radar chain consists of signal processing algorithms that require high computing power. Multi-processor and multi-core parallel embedded machines are one of the solutions to meet real-time constraints of many radar applications. In this paper, we proposed efficient and scalable parallelization methods of the Pulse-Doppler radar signal processing chain. First, we evaluated Open Multi Processing (OpenMP) to identify its best scheduling technique in order to exploit efficiently the available computing cores. Then, we have proposed new parallel and scalable approaches based on direct memory access (DMA) and inter-processor communication (IPC) techniques, combined with the best OpenMP scheduling method to accelerate radar signal processing chain. To prove the scalability of our proposed parallel approaches, two radar use cases with different real-time and memory constraints have been experienced. We used the eight cores C6678 digital signal processor (DSP) as a target for all our implementations. The obtained results show an overall parallel efficiency of 95%, which is better than the best state-of-the-art implementations.

In this paper, we present a processing method to detect millimeter interlayer debondings from Ground Penetrating Radar (GPR) B-scan images. The method is matched to carry out rapid debonding detection at the operational level. A machine learning based outlier-detection strategy namely, One-class Support Vector Machines (OCSVM) is proposed to detect A-scan data vectors which differ from a reference data set collected over a known healthy pavement area. OCSVM is tested on both simulated and experimental data representing GPR data over various artificial millimetric debondings at 2.6 GHz and 4.2 GHz from respectively ground-coupled and air-coupled radar configurations. The experimental data were collected at the Accelerated Pavement Test site located in the Nantes campus of Universite Gustave Eiffel. The simulated models on the other hand were generated using a numerical EM solver based on Finite Difference Time Domain (FDTD) method namely, GprMax. Simulation tests allow to conduct sensitivity analysis to determine the robustness of the detection method at various signal-to-noise ratios (10 dB to 60 dB). The proposed OCSVM method demonstrated high performance on both simulated and experimental data to detect thin interlayer debondings over various GPR configurations.

This paper presented a novel design of thin and wide-band electromagnetic absorber. The proposed structure was composed of a bottom metal plate, a homogeneous resistance layer (HRL), a dielectric spacer, and a periodic resistance layer (PRL) at the top made of the Jerusalem cross structure, and the overall thickness of the absorber was only 5.055 mm. The simulation result exhibited that more than 90% absorption has been achieved in 2–40 GHz, corresponding to the fractional bandwidth of 181% under the normal incidence. The measured and the simulated results showed an agreement. The absorber had a good potential application in the field of microwave absorption.

Deep Convolutional Neural Networks are finding their way into modern machine learning tasks and proved themselves to become one of the best contenders for future development in the field. Several proposed methods in image segmentation and classification problems are giving us satisfactory results and could even perform better than humans in image recognition tasks. But also at the cost of their performance, they also require a huge amount of images for training and huge amount of computing power and time that makes them unrealistic in some situations where obtaining a large dataset is not feasible. In this work, an attempt is made for segmentation of Synthetic Aperture Radar (SAR) images which are not usually abundant enough for training, and are heavily affected by a kind of multiplicative noise called speckle noise. For the segmentation task, pre-defined filters are first applied to the images and are fed to hybrid CNN that is resulted from the concept of Inception and U-Net. The outcome of our proposed method has been examined for their effectiveness of application in a complete set of SAR images that are not used for training. The accuracy has also been compared with the manually annotated SAR images.

High-resolution range profile (HRRP) data often has imbalanced and open-ended distribution in realistic radar target recognition task, requiring the recognition system classify targets among majority and minority classes and detect unseen targets accurately. In this paper, we propose a novel algorithm for open and imbalanced HRPP recognition tasks, by learning from realistic data distribution and optimizing the accuracy over the majority, minority and open classes. This method maps the HRRP data to the latent feature space, enhances the feature by sharing knowledge from majority to minority class, and provides a scalar indicating the familiarity to known classes. Dual-attention is developed to provide strong discriminative feature representation. An angular penalty is employed in the loss function to optimize the intra-class similarity and inter-class variability. Experiments on measured data prove that the proposed algorithm outperforms other existing methods under both close and open settings with accuracy increased significantly, and more discriminative representation. This study provides a promising and effective approach for open and imbalanced HRRP target recognition.

Spiking neural networks (SNNs) have recently gained large interest for edge-AI applications due to their low latency and ultra-low energy consumption. Unlike DNNs, SNNs communicate information using spike trains. As the derivative of spike trains are highly ill-defined, the use of surrogate gradients has been proposed as an efficient method for training SNNs. Still, the lack of open-source SNN softwares and the limited range of demonstrated SNN applications slows down a wider SNN adoption. We release our ConvSNN framework, demonstrating the novel applicability of quantized-weight SNNs for radar gesture recognition. Our framework will facilitate future research in the SNN area.

Based on cloud-probe data and airborne Ka-band cloud radar data collected in Baoding on 5 August 2018, the microphysical structural characteristics of cumulus (Cu) cloud at the precipitation stage were investigated. The cloud droplets in the Cu cloud were found to be significantly larger than those in stratiform (STF) cloud. In the Cu cloud, most cloud particles were between 7 and 10 μm in diameter, while in the STF cloud the majority of cloud particles grew no larger than 2 μm. The sensitivity of cloud properties to aerosols varied with height. The cloud droplet effective radius showed a negative relationship with the aerosol number concentration (Na) in the cloud planetary boundary layer (PBL) and upper layer above the PBL. However, the cloud droplet concentration (Nc) varied little with decreased Na in the high liquid water content region above 1500m. High Na values of between 300 and 1853 cm−3 were found in the PBL, and the maximum Na was sampled near the surface in August in the Hebei region, which was lower than that in autumn and winter. High radar reflectivity corresponded to large FCDP (fast cloud droplet probe) particle concentrations and small aerosol particle concentrations, and vice versa for low radar reflectivity. Strong updrafts in the Cu cloud increased the peak radius and Nc, and broadened cloud droplet spectrum; lower air temperature was favorable for particle condensational growth and produced larger droplets.

Objective Non-contact continuous biometric identification system based on the heartbeat signals has attracted more attention due to its privacy-friendly properties. Current methods, however, mostly focused on the traditional machine learning methods under the close-set condition. This paper aims to investigate the feasibility of using the cardiac radar heartbeat signals and deep learning techniques to identify person in the open-set environment. Methods A novel dipole deep learning model (DDLM) was proposed for the open-set person identification problem without heartbeat segmentation and feature engineering. The normalized heartbeat samples with time duration of 5 s were used as input and encoded into the feature space, where the encoded features of the same person cluster closely around the corresponding negative pole and repel far from positive pole, and those of different persons separate loosely from each other. Finally, threshold on the distance from the features to the dipoles in the feature space was set for each known identity. Results Extensive experiments conducted on a public dataset of clinically recorded vital signs indicate that:(1) The proposed model shows high stability under close-set condition with an accuracy higher than 99 % with 30 subjects.(2) The accuracy and the F1-score attain 93.42 % and 93.57 % under the open-set condition with the maximum openness of 29.3 %, respectively. Conclusion The proposed model shows high effectiveness in person identification using heartbeat signals. The DDLM outperforms most of the existing methods under the close-set condition. And the DDLM shows a promising future for person identification in open-set environment.

In this paper, we developed a new design of multiband perfect metamaterials absorber (MTMA) to be operated in the frequency range of 8–12 GHz. Proposed structure consists of a combination of three basic structures which is based on Closed Circular Ring Resonators (CCRR). A dielectric substrate is chosen as FR-4 material of 1.6 mm thickness and 22 × 22 mm size was used. The proposed (MTMA) operated at six different resonance frequencies 8.3 GHz, 9.2 GHz, 9.88 GHz, 10.66 GHz, 11.16 GHz, and 11.85 GHz with the absorption of 100 %, 100%, 98%, 91%, 98%, and 98% respectively. Negative parameters such as permittivity ɛ permeability ( μ ) reflective index ( n ) , and impedance ( z ) were extracted and presented. Results indicated that the MTM absorber was found to extremely independent on the angles of polarization of the incident waves and the absorber was almost insensitive to the incident angle (0° – 90°) for transverse electric (TE) or transverse magnetic (TM) modes. The working principle of the structure is analyzed and illustrated through input impedance, surface current, and the electric field of the structure. The designed absorber was fabricated and experimentally measured in microwave laboratory that shows a good agreement with simulation results. According to the obtained results, the proposed metamaterials absorber is a good candidate for the military RADAR applications.

PURPOSE/OBJECTIVE(S) To assess the efficacy of a Risk-Adapted Ablative Radiotherapy (RAdAR) approach, after intensive induction chemotherapy, in patients with locally advanced pancreatic cancer (LAPC). MATERIALS/METHODS Patients with LAPC treated with RAdAR from January 2017 to December 2019 were included in this observational study. The RAdAR approach consisted of an anatomy- and simultaneous integrated boost (SIB)-based dose prescription strategy. RAdAR was delivered with stereotactic ablative radiation therapy (SAbR), administering 30 Gy in 5 fractions to the tumor volume (PTVt) and 50 Gy SIB (BED10 100 Gy) to the vascular involvement. If SAbR was not feasible (tumor dimension ³ 6 cm, extensive nodal spread, duodenal infiltration), an (hypo-)fractionated ablative radiotherapy (HART) prescribing 50.4 Gy in 28 fractions to the PTVt, with a vascular SIB of 78.4 Gy (BED10 100 Gy), was adopted. Primary end points were freedom from local progression (FFLP), overall survival (OS), and progression-free survival (PFS). RESULTS Sixty-four LAPC patients were included. Induction chemotherapy consisted of gemcitabine/nab-paclitaxel in 60.9% and FOLFIRINOX in 39.1% of cases. SAbR was used in 52 (81.2%) patients, and HART in 12 (18.8%). After RAdAR, surgery was performed in 17 (26.6%) patients. Median follow-up was 16.1 months (22.9 months among patients still alive at the last observation). Overall local control (LC) rate was 78.1%, with no difference between resected and non-resected patients (2-year FFLP 75.3% vs 56.4%; P = 0.112). Median OS and PFS were 29.7 months and 8.7 months, respectively, for the entire cohort. Resected patients had a better median OS (not reached versus 26.1 months; P = 0.0001) and PFS (19 versus 5.6 months; P < 0.0001) compared to non-resected patients. In non-resected patients, no significant difference was found between SAbR and HART for median FFLP (28.1 versus 18.5 months; P = 0.614), OS (27.4 versus 25.3 months; P = 0.624), and PFS (5.7 versus 3.4 months; P = 0.486). One patient (1.6%) experienced acute grade 4 gastro-intestinal bleeding. No other acute or late grade ≥ 3 toxicities were observed. CONCLUSION The RAdAR approach, following intensive induction chemotherapy, is an effective radiation treatment strategy for selected LAPC patients, representing a promising therapeutic option in a multimodality treatment regimen. Further studies are needed to corroborate these results.

Automatically recognizing intra-pulse modulation of radar signals is a significant survival technique in electronic intelligence systems. To avoid the dependence on feature selection and realize the intelligent intra-pulse modulation recognition of various radar signals under low signal-to-noise ratios (SNRs), this paper develops a novel intra-pulse modulation recognition method based on the high-order spectrums of radar signals. Automatic soft thresholding is implemented in the deep residual network to adaptively eliminate redundant information in the process of feature learning and improve the learning effect of valuable features in distribution images of corresponding third-order spectrums. The extensive simulations compared with the other four methods further reveal the excellent classification performance of the proposed method. The proposed approach still achieves an overall probability of successful recognition of 93.5% for eight kinds of modulation signals, even when the SNR is just −8 dB. Outstanding performance proves the superiority and robustness of the proposed method.

Both polymeric electrochromic and radar absorption materials have been considered as science hotspots over the past few decades. However, integrating both properties (electrochromic and radar absorption) in a single device has not been addressed to fabricate a multifunctional device so far. Thus, here, we endeavor to shed light on this uncharted area by presenting a new smart ternary nanocomposite (polythiophene-based/RGO-Fe3O4) which can provide both electrochromic and microwave absorbent features simultaneously. In this regard, we synthesized the binary Fe3O4-RGO nanocomposites with different contents of Fe3O4 nanoparticles. Then, this binary nanocomposite is utilized in combination with a polythiophene-based conductive polymer as a tricolor (RGB) organic electrochromic material to synthesize the smart ternary nanocomposite through both in-situ polymerization and physical mixing methods. The ternary nanocomposites containing different weight fractions of Fe3O4-RGO nanoparticles were used to assemble the rigid and flexible ECD devices using flat and fiber substrates such as glass-indium tin oxide (glass-ITO), polyethylene terephthalate-indium tin oxide (PET-ITO) and stainless steel wire (SSW). By applying a voltage of −2.1 to 2.1 v, a reversible color-switching ability was observed between green color (oxidized state), and red color (reduced state) for all ECDs. The average switching time and color memory for the SSW cell were determined to be about 3 s and 155 min, and for both glass-ITO and PET-ITO cells to be 2 s and 210 min, respectively. Additionally, the microwave absorption properties of the prepared ternary nanocomposites were characterized by measuring the electromagnetic parameters such as permittivity, permeability, and reflection loss. A ternary nanocomposite with 3 mm thickness filled with 50 wt% Fe3O4-RGO exhibited a maximum reflection loss as high as −30.2 dB (92% absorption) at 12.3 GHz (in the range of x-bond) with 1.1 GHz bandwidth.

Here we report on a modulated sub-nanosecond lidar-radar for underwater target detection. The modulated laser source is consisted of a 1064 nm master oscillator power amplifier (MOPA) and a frequency-doubling module, outputting pulse energy of 178.1 mJ at 1064 nm with 0.76 ns pulse duration and 87.6 mJ at 532 nm. By using a Fabry–Perot cavity containing a KTP crystal and two reflectors, the 532 nm laser pulse is modulated into the burst mode with a repetition rate of 500 MHz. A streak tube camera is used as the signal receiving apparatus. An underwater target detection experiment has been carried out. The results indicate that this lidar-radar could emit a stable and powerful modulated signal, which greatly increases the underwater detection range. The cooperation of the streak tube imaging and modulated laser source for underwater target detection experiments is the first attempt in this field. The clear 3-D and 4-D images of the target underwater across 20 m are obtained in the experiments and the spatial resolution of 9 mm can be achieved.

Perennial commodity crops, such as coffee, often play a large role globally in agricultural markets and supply chains and locally in livelihoods, poverty reduction, and biodiversity. Yet, the production of spatial information on these crops are often overlooked in favor of annual food crops. Remote sensing detection of coffee faces a particular set of challenges due to persistent cloud cover in the tropical “coffee belt,” hilly topography in coffee growing regions, diversity of coffee growing systems, and spectral similarity to other tree crops and agricultural land. Looking at the major coffee growing region in Dak Lak, Vietnam, we integrate multi-temporal 10 m optical Sentinel-2 and Sentinel-1 SAR data in order to map three coffee production systems: i) open-canopy sun coffee, ii) intercropped and other shaded coffee and iii) newly planted or young coffee. Leveraging Google Earth Engine (GEE), we compute five sets of features in order to best enhance separability between coffee and other land cover and within coffee production systems. The features include Sentinel-2 dry and wet season composites, Sentinel-1 texture features, Sentinel-1 spatiotemporal metrics, and topographic features. Using a random forest classification algorithm, we produce a 9-class land cover map including our three coffee production classes and a binary coffee/non-coffee map. The binary map has an overall accuracy of 89% and the three coffee production systems have user accuracies of 65, 56, 71% for sun coffee, intercropped coffee and newly planted coffee, respectively. This is a first effort at large-scale distinction of within-crop production styles and has implications across many applications. The binary coffee map can be used as a high-resolution crop mask, whereas the detailed land cover map can inform monitoring of deforestation dynamics, biodiversity, sustainability certification and implementation of climate adaptation strategies. This work offers a scalable approach to integrating optical and radar Sentinel data for production of spatially explicit agricultural information and contributes particularly to tree crop and agroforestry mapping, which often is overlooked in between agricultural and forestry sciences.

Breathing rate monitoring for a long period provides a valuable indicator about human health and an early warning of possible disasters on the well-being. To gain a comfortable and adequate measurement without restricting the person's privacy, this study investigates a novel non-contact-based solution using ultra wide-band (UWB) impulse radar. We propose a fast and low-cost method for breathing events detection and respiration rate (RR) tracking. This novel approach consists of three main parts: vital signs (VS) patterns extraction based on the Generalized Goertzel Algorithm (GGA) avoiding the prior knowledge of the person's location and overcoming the effect of the clutter. RR tracking built on the complex adaptive notch filter (CANF) with a variable step size to obtain a robust estimation under the presence of sparse random body movement (SRBM) and fast tracking of the breath variability. Finally, we introduce a control chart of the breathing process using the double exponential moving average (DEMA) in order to detect the apnea events. The proposed method is evaluated using two sets of experiments, from mechanical vibrating structure to human subject. The results demonstrate that the proposed system is a promising and effective solution and provides wide usability for continuous monitoring applications. Moreover, it outperforms the state-of-the-art, in terms of computation cost and space complexity.

Changes in the frequency and magnitude of extreme hydrological events (floods and droughts) are one of the most significance consequences of the climate change and their monitoring is therefore essential in many fields. However, the number of the global in situ discharge measurements has been declining since the eighties and the global supply of available data is often restricted in many parts of world. All these reasons have fostered in the last decades the exploitation of new technologies, such as satellite data, in the estimation of river discharge. In this study, satellite altimetry water level data are used to set-up a consistent and up-to-date river discharge dataset by means of rating curves for different sites across the world. Long continuous river discharge time series so obtained are used to perform a trend analysis on extreme hydrological events. Specifically, peak-over threshold values are extracted from the discharge time series, as proxy variables of independent extreme events, and the trend analysis is then carried out to identify changes in the magnitude of extreme events. The analysis permits to identify areas of the world prone to floods, like the West Africa, or drought so that appropriate actions for disaster risk mitigation and continuous improvement in disaster preparedness, response and recovery practices can be adopted.

Stealth aircraft use Stealth Technology in order to evade recognition by utilizing mix of highlights with the goal that Aircraft can invade with RADAR and furthermore decrease the deceivability in “radio frequency (RF) range”, sound and infrared visual. During the most recent decades, stealth technology has demonstrated to be one of the best methodologies to the extent the undertaking to escape radar systems is concerned. Stealth intends to evade the recognition or attempt to cover up, for planes, avoiding radar implied stealth. The idea utilized in the stealth innovation isn’t so difficult. It utilizes the guideline to retain and mirror the radar waves. Airplane redirects the radar waves in different ways and limits the radar waves which return back to radar. Another marvel which permits the air ship maintain completely the incoming radar wavelengths to inhabit the electromagnetic reincorporated energy in various way. The structure and material used to manufacture air ship choose the degree of stealth accomplish via air ship. The thought for the radar radio wire to release Radio energy that is then returned through any item that occurs to recognize. The time taken for the reflection to come is measure by the radar radio wire and it can tell how far off an object is.

Producing high-resolution urban land cover maps is essential for decision-making and urban management. In this regard, Synthetic Aperture Radar (SAR), especially Polarimetric SAR (PolSAR), has served as a valuable data source to fulfill this task. Direct conversion of low-level features into high-level Land Cover (LC) concept may reduce the final classification accuracy. Therefore, mid-level representation models, such as Bag of Visual Words (BOVW), were employed to resolve the existing semantic gap challenge. In this paper, a Segment-based BOVW (Seg-BOVW) model was developed for urban land cover classification using PolSAR data. To this end, two PolSAR data over San Francisco Bay (SFB) and Flevoland (FL) acquired by RADARSAT-2 were employed to comprehensively evaluate the Seg-BOVW model's performance. First, to exploit the full potential of PolSAR data, 169 low-level features in four categories: (1) original, (2) polarimetric, (3) texture, and (4) decomposition features were extracted. Afterward, a Multi-Objective Genetic Algorithm (MOGA) was implemented to investigate the importance of low-level features for urban land cover mapping. This step resulted in selecting 14 and 6 low-level features, as the high contributing features, for SFB and FL datasets, respectively. The Seg-BOVW model achieved significant overall accuracies of 96.02% and 98.82% for SFB and FL, respectively, indicating the high potential of the proposed method for urban land cover classification. Furthermore, a comparison with other well-known algorithms of Support Vector Machine (SVM), Artificial Neural Networks (ANN), and Convolutional Neural Networks (CNN) was made, suggesting the capability of the Seg-BOVW model to improve the urban land cover classification results. Finally, the Seg-BOVW model was tested with two other PolSAR datasets acquired with different sensors over SFB to examine its applicability with different datasets.

PURPOSE To compare medial elbow stress during squatting throwing, standing throwing, and maximum effort throwing and to analyze its relationships with throwing variables and the hip joint internal rotation (IR) range of motion (ROM) in youth baseball catchers. METHODS Forty-five youth baseball catchers were recruited. Exclusion criteria included pain during throwing, history of surgery, or upper-extremity injuries within 12 months. Catchers performed 5 throws for each of the following throwing conditions: squatting throwing, standing throwing, and maximum effort throwing. Catchers were outfitted with a wearable sensor sleeve that recorded the medial elbow torque, arm speed, and shoulder rotation. Ball velocity was measured using a radar gun. RESULTS Final analysis included 42 catchers (mean age, 11.5 ± 1.2 years; mean catcher experience, 16.7 ± 10.3 months). The medial elbow torque was significantly greater during maximum effort throwing (19.0 ± 5.2 Nm) than during squatting throwing (14.3 ± 5.4 Nm) and standing throwing (15.7 ± 4.7 Nm) (both P < .001); there was no significant difference between squatting throwing and standing throwing. Multiple linear regression analysis showed that the ball velocity was significantly positively associated with the medial elbow torque during squatting throwing and standing throwing (P = .023 and P = .029, respectively). During squatting throwing, the hip IR ROM on the throwing side was significantly negatively associated with the medial elbow torque (P = .011). CONCLUSIONS The medial elbow torque during squatting throwing was equivalent to that during standing throwing. Increased medial elbow torque was associated with greater ball velocity during squatting throwing and standing throwing and with decreasing hip joint IR ROM on the throwing side during squatting throwing. Evaluation and stretching of the hip joint IR ROM on the throwing side may be important to decrease elbow stress during squatting throwing in youth baseball catchers. LEVEL OF EVIDENCE Level II, prospective comparative observational trial.

Landslide susceptibility analysis is vital for understanding precursors, associated risk, and early rescue operations. In the past, numerous attempts have been made to develop landslide susceptibility maps using various tools. This research work aims to generate landslide susceptibility maps (LSM) by developing a deep multi-layer perceptron network named as DMLP-LSM and to demonstrate it over Srinagar-Rudraprayag region, a landslide prone region recognized by the Geological Survey of India (GSI). Initially, training datasets are generated using the analytical hierarchy process (AHP) enabled multi-criteria analysis, which involves information from various landslide causative factors. In a first, one dimensional line of sight (1D LOS) velocity, one of the critical inputs and prominent indicator for landslide susceptibility, is generated by processing 60 Sentinel-1A scenes using multi-temporal interferometric synthetic aperture radar (MT-InSAR) algorithm. The 1D LOS velocity map, generally used for quantitative monitoring slow moving landslides, is used here as a thematic layer. The thematic layers and LSMs were fed to the proposed DMLP-LSM network as input and output, respectively. The developed model reached an accuracy of 91.47% with precision, specificity and ROC_AUC score as 92.11%, 95.55% and 0.9072, respectively. It further identified nearly 15 square km area as susceptible to landslide. The efficacy of the proposed model is compared with two well-known machine learning models for LSM preparation, namely support vector machine (SVM) and random forest (RF). In terms of accuracy, DMLP-LSM and RF had similar performance, but better than SVM. The velocity maps derived from MT-InSAR processing played a prominent role in the generation of LSM, supporting model training and performance. The proposed model achieved highest precision and specificity, and its performance would also improve LSM preparation when using larger datasets. The obtained results are promising, and the trained model can help in the rapid generation of LSMs.

This study compares the utility of multifrequency SAR and Optical multispectral data for land-cover classification of Mumbai city and its nearby regions with a special focus on water body mapping. The L-band ALOS-2 PALSAR-2, X-band TerraSAR-X, C-band RISAT-1, and Sentinel-2 datasets have been used in this work. This work is done as a retrospective study for the dual-frequency L and S-band NASA-ISRO Synthetic Aperture Radar (NISAR) mission. The ALOS-2 PALSAR-2 data has been pre-processed before the implementation of machine learning algorithms for image segmentation. Multi-looking is performed on ALOS-2 PALSAR-2 data to generate square pixels of size 5.78 m and then target decomposition is applied to generate a false-color composite RGB image. While in the case of TerraSAR-X and RISAT-1 datasets, no multi-looking was performed and direct target decomposition was applied to generate false-color composite RGB images. Similarly, for the optical dataset that has a resolution of 10 m, a true color composite, and a false color composite RGB image are generated. For the comparative study between ALOS-2 PALSAR-2 and Sentinel-2 dataset, the RGB images are divided into smaller chunks of size 500*500 pixels each to create a training and testing dataset. Ten image patches were taken from the large dataset, out of which eight patches were used to train the machine learning models Random Forest (RF), K Nearest Neighbor (KNN), and Support Vector Machine (SVM), and two patches were kept for testing and validation purpose. For training the machine learning models, feature vectors are generated using the Gabor filter, Scharr filter, Gaussian filter, and Median filter. For patch 1, the mIOU for true-color composite based Optical image varies from 0.2323 to 0.2866 with the RF classifier performing the best and the mIOU for false-color composite based Optical image varies from 0.4130 to 0.4941 with the RF classifier performing the best while for ALOS-2 PALSAR-2 data, the mIOU varies from 0.4033 to 0.4663 with the RF classifier outperforming the KNN and the SVM classifiers. For patch 2, the mIOU for true-color composite based Optical data varies from 0.3451 to 0.4517 with KNN performing the best and the mIOU for false-color composite based Optical image varies from 0.5156 to 0.5832 with the RF classifier performing the best while for ALOS-2 PALSAR-2 data, the mIOU varies from 0.4600 to 0.5178 with the RF classifier outperforming the KNN and the SVM classifiers. The gap between the performance of ALOS-2 PALSAR-2 data and Sentinel-2 optical data is observed when the IOU of water class is compared, with IOUw for the true-color composite based optical image at a maximum of 0.2525 and for false-color composite based optical image at a maximum of 0.7366 while for ALOS-2 PALSAR-2 data a maximum IOUw of 0.7948 is achieved. The better performance of SAR data as compared to true-color composite based optical image data is due to the misclassification of ground and water classes into urban and forest in the case of the true-color composite based optical dataset which can be attributed to the high similarity between water and forest classes in the case of true-color composite based optical data whereas both these classes are easily separable in case of SAR data. This issue is however resolved by using the false color composite based optical image dataset for the classification task which performs slightly better than ALOS-2 PALSAR-2 data in the overall classification task. However, the SAR data works best in water body detection as notable from the high IOU for water class in the case of SAR data. In addition to the comparative analysis between Sentinel-2 optical and ALOS-2 PALSAR-2 data, land-cover classification has been performed on X-band TerraSAR-X and C-band RISAT-1 data on a single patch and it has been found that the RF classifier performs the best, recording the mIOU 0.5815 for X-band TerraSAR-X data, mIOU of 0.4031 for the C-band RISAT-1 data, and mIOU of 0.6153 for the L-band ALOS-2 data.

Regional as well as global cropland database at a high spatial resolution like 10 m is not available. This study aims to integrate the Sentinel-1 synthetic aperture radar (SAR) data and Sentinel-2 optical data in a more robust way to map the croplands at 10 m spatial resolution. It also examines the effectiveness of different derived products from the SAR and multispectral data. The VH and VV polarizations of the SAR data were used to generate the 80 textural parameters at four different kernel sizes. The multispectral bands of the optical data were used to create nine different spectral indices. The SAR and optical data were classified separately and integrated with the other derived products, using the random forest classifier, to identify the most potential way to use the two datasets for LULC classification. The classification accuracy was the least when only the Sentinel 1 data was used, with an overall accuracy of 59.48% and a kappa value of 0.51. The best accuracy (97.94% overall accuracy and 0.98 kappa value) was achieved by combining the Sentinel 1 and 2 data and their optimal textures and spectral indices. Overall, this study suggests the suitability of the SAR data along with optical data to map the croplands with higher accuracy.

A country’s food needs mainly depend upon its agriculture resources and require reliable information related to crop health, distribution, and acreage estimation to manage and monitor resources to implement a sustainable agricultural system. Different methodologies have been used to collect this information. However, the availability of earth resource satellite data with improved spatial, spectral, and temporal resolutions, such as the European Space Agency's (ESA) Copernicus program satellites Sentinel-1 (S1) and Sentinel-2 (S2), are creating more practicability to generate crop type maps. S1 and S2, both operating in a constellation of twin satellites, carry a C-band Synthetic Aperture Radar (SAR) and Multispectral Instrument (MSI). Vertical transmit and horizontal receive (VH), and vertical transmit and vertical receive (VV) channels of S1 were used to exploit the temporal backscatter of crops present in the study area. In this research, a machine learning random forest classification algorithm is used for accurate crop type mapping through the combination of SAR (S1) and optical (S2) time-series data. Random forest classifier has produced considerable improved accuracies of crop type mapping in previous studies as it uses ensemble decision trees trained on sample data which permit the vote in favor of the most popular land use/ land cover class. The key objectives of this study are to investigate the classification accuracies for different data combinations. Three plots of data are tested (i) S1 (ii) S2 (iii) S1 & S2. A combination of SAR and optical data turn out with the best overall accuracy of 97% and a kappa coefficient of 0.97. Space-borne SAR and optical data add a new aspect of crop type mapping, which increases the classification accuracy by including valuable parameters and beating the drawbacks of each other. By comparing the results, it can be concluded that combining all-weather accessible SAR and spectrally rich optical data accomplished more accurate outcomes. It would be an imperative advance for the future endeavor to estimate crop biomass and biophysical parameters.

In this paper, we consider the problem of fusion of synthetic aperture radar (SAR) images from spaceborne and airborne sensors and investigate its applications to inshore ship target detection. Existing SAR image fusion methods mainly focus on image denoising or texture enhancement, but show limited improvement of target-to-clutter ratios (TCRs) in composite images and lead to deteriorated target detection performance. To address this issue, we propose a new method for the fusion of spaceborne and airborne SAR images based on the target proposal and the copula theory (TPCT). In TPCT, target and clutter correspondence between different images are exploited to improve the TCRs of composite images. TPCT consists of three steps. First, target proposals are extracted from spaceborne and airborne SAR images and then fused to enhance the common ship target areas therein. Second, a new method to construct the joint probability density function (PDF) of clutter in spaceborne and airborne SAR images is presented to model the statistical dependence of clutter therein based on the copula theory. This copula-based joint PDF is used to suppress the clutter areas remained in the intersection of target proposals. Third, clues from the intersection of target proposals and the copula-based joint PDF of clutter are fused by the Hadamard product to generate the composite image with enhanced ship targets and the suppressed clutter. Experimental results based on measured spaceborne and airborne SAR data show that the proposed TPCT fusion method leads to higher TCRs of composite images and better performance in the ship detection task than other commonly used image fusion methods.

Identifying a spectral frequency band with abundant fault information from spectral coherence is essential for improved envelope spectrum-based bearing diagnosis. Both blind features and targeted features have been employed to distinguish informative spectral frequency band of spectral coherence. However, how to select appropriate feature to correctly discriminate the optimal frequency band of spectral coherence in different scenarios is problematic. In this study, a new targeted feature is presented to quantify the signal-to-noise ratio in narrow frequency bands of spectral coherence, and further a method based on the proposed feature is developed to distinguish an optimal spectral frequency band of spectral coherence for bearing diagnostics. The efficiency of the developed method, typical blind feature-based methods and typical targeted feature-based methods in identifying the defect-sensitive frequency band of spectral coherence and bearing fault diagnosis is validated and compared using simulated signals with different interference noises and bearing experimental signals. The advantages and limitations of typical blind and targeted feature-based methods in different scenarios are summarized to guide the application. The results demonstrate that the developed targeted feature can efficiently evaluate bearing failure information in the cyclic frequency domain, and the presented approach can accurately discriminate the failure-related spectral frequency band of spectral coherence and detect different bearing faults compared with the methods based on the state-of-the-art features.

With the accessibility of Synthetic Aperture Radar (SAR) data and evolution of deep learning, several experiments have been performed in the recent years to actively monitor human activities at sea. Due to the flexible network configuration and complex nature, models based on Convolutional Neural Networks have been able to boost the performance of maritime target classification systems. The novel idea behind these techniques is to enhance the efficiency and automation in the process. However, the limited availability of labeled data and lower resolution of Sentinel-1 (SAR) images have constrained the accuracy of classification models based on the deep learning methods. Besides, it is observed that image processing techniques have resulted in the loss of information present in the images. To overcome these issues, a new method has been developed in the present work that integrates transfer learning techniques (referred to as ResNet50) with additional scale-variant features extracted before image processing. To test our model, SAR and Automatic Identification System (AIS) data were collected from 3 locations: Jawaharlal Nehru (JN) Port (India), Singapore Port (Singapore), and Shanghai-Zhoushan Port (China). A labeling process was done by using the SAR-AIS Neighbourhood integration. Experimental work showed that the proposed architecture of ResNet50 model synergized with extracted scale variant features provides promising results with the classification accuracy of 82.35% and F score of 0.82 on the test dataset, when no models have been developed and trained on Sentinel-1 20 m resolution data for given ship classes.

Unlike the existing change detection approaches based on the multispectral (MS) image and synthetic aperture radar (SAR) image datasets, a novel unsupervised hyperspectral change detection (UHCD) framework is proposed in this paper. The UHCD framework is designed for hyperspectral images with high dimensions and low availability. This framework consists of two modules: the spectral mapping with adversarial learning and the discriminant analysis with spatial attribute optimization. In comparison with other advanced change detection methods, the proposed framework possesses three distinctive properties: (1) The unsupervised spectral mapping is leveraged to exploit underlying spectral features without the requirement of pseudo-training datasets in the change detection task; (2) We introduce spectral constraint loss into reconstruction space and adversarial loss into latent space to enhance the quality of the features extracted by the spectral mapping network; (3) Spatial attribute optimization uses the spatial correlation to further improve the performance of the proposed UHCD method. The experimental results on two real datasets show that the proposed UHCD achieves competitive performance.

Objectives To examine the efficacy of “slow” signs and patroller presence at “slow” signs to reduce speeds of snowsports participants, compared to a condition where no sign or patroller are present, independent of other factors that may contribute to skier slowing (such as prior knowledge, trail convergence, etc.). Design and methods Snowsports participant speeds were measured on “more difficult” trails using a radar gun at two ski areas with: (1) no-sign — the usual condition for the trail, and (2) slow-sign — a large “slow” sign was posted in the middle of the trail. At one ski area, a third condition was also tested: (3) slow + patroller — a ski patroller stood at the slow sign. Participant equipment type and estimated ability were also recorded. Results At one ski area, there was no significant difference in speed between conditions. At the second ski area, the differences in mean (SD) speeds were small but significant for the no-sign, slow-sign, and slow + patroller conditions: 10.9 (3.0), 10.3 (2.9), and 9.8 (2.6) m/s. Effects were driven by non-beginner skiers; on average, beginner skiers and all snowboarders were slower than non-beginner skiers and did not adjust their speed in response to the signage conditions. Conclusions Reductions in speed for the slow-sign and slow + patroller conditions compared to the no-sign condition were small (0.5 and 1.1 m/s) compared to the variation in chosen skier speed reported in other studies. The small differences in chosen speeds make it unlikely that slow sign and patroller presence alone would produce meaningful reductions in collision likelihood or severity of impacts.