Search

To deal with the complexities of decision-making for unmanned aerial vehicles (UAVs) in denial environments, this paper applies deep reinforcement learning algorithms to search and rescue (SAR) tasks. It proposes a two-stage target search and tracking method for UAVs based on deep reinforcement learning, which divides SAR tasks into search and tracking stages, and the controllers for each stage are trained based on the proposed deep deterministic policy gradient with three critic networks (DDPG-3C) algorithm. Simulation experiments are carried out to evaluate the performance of each stage in a two-dimensional rectangular SAR scenario, including search, tracking, and the integrated whole stage. The experimental results show that the proposed DDPG-3C model can effectively alleviate the overestimation problem, and hence results in a faster convergence and improved performance during both the search and tracking stages. Additionally, the two-stage target search and tracking method outperforms the traditional single-stage approach, leading to a more efficient and effective decision-making ability in SAR tasks.

Due to hardware technology limitations, satellite sensors are unable to capture images with high temporal, spatial, and spectral resolutions simultaneously. However, the Gaofen-1 satellite overcomes this challenge by incorporating 2-meter panchromatic, 8-meter multispectral, and 16-meter wide-field cameras, allowing for the integration of images from these sensors. To address this issue, we propose a study on the spatio-temporal-spectral fusion method for Gaofen-1 images, aiming to achieve more comprehensive structures. Inspired by the diffusion model, which learns the data distribution of the target image, we propose a new network utilizing an enhanced diffusion framework. The network incorporates both structural and spectral constraints to guide the fusion process. This work represents the first application of the diffusion model to spatio-temporal-spectral fusion, specifically synthesizing the 2-meter multispectral images with dense temporal resolution. To assess fusion quality, we have developed a benchmark dataset. During the validation stage, we evaluate the radiometric deviation, structural similarity, and spectral fidelity between the fused 2-meter multispectral images and the reference images. Both visual and quantitative assessments demonstrate that our newly proposed method work well for the Gaofen-1 fusion.

Fixed-wing Vertical Takeoff and Landing (VTOL) drones have been widely researched and applied because they combine the advantages of both rotorcraft and fixed-wing drones. However, the research on the transition mode of this type of drone has mainly focused on completing the process quickly and stably, and the application potential of this mode has not been given much attention. The objective of this paper is to routinize the transition mode of compound VTOL drones, i.e., this mode works continuously for a longer period of time as a third commonly used mode besides multi-rotor and fixed-wing modes, which is referred to as the hybrid mode. For this purpose, we perform detailed dynamics modeling of the drone in this mode and use saturated PID controllers to control the altitude, velocity, and attitude of the drone. In addition, for more stable altitude control in hybrid mode, we identify the relevant parameters for the lift of the fixed-wings and the thrust of the actuators. Simulation and experimental results show that the designed control method can effectively control the compound VTOL drone in hybrid mode. Moreover, it is proven that flight in hybrid mode can reduce the flight energy consumption to some extent.

Land use recognition from multispectral satellite images is fundamentally critical for geological applications, but the results are not satisfied. The scale dimension of current multiscale learning is too coarse to account for rich scales in multispectral images, and pixel-wise classification tends to produce “salt-and-pepper” labels due to possible misclassification in heterogeneous regions. In this article, these issues are addressed by proposing a new pixel-wise classification model with finer scales for convolutional neural networks. The model is designed to extract multiscale contextual information using multiscale networks at a fine-grained level, addressing the issue of insufficient multiscale learning for classification. Furthermore, a small-scale segmentation-combination method is introduced as a postprocessing solution to smooth fragmented classification results. The proposed method is tested on GF-1, GF-2, DEIMOS-2, GeoEye-1, and Sentinel-2 satellite images, and compared with six neural-network-based algorithms. The results demonstrate the effectiveness of the proposed model in finding objects of large scale difference, improving classification accuracy, and reducing classified fragments. The discussion also illustrates that convolutional neural networks and pixel-wise inference are more practical than transformer and patch-wise recognition.

A flywheel energy storage system (FESS) achieves energy conversion through a permanent magnet synchronous machine (PMSM). The PMSM in a FESS requires low current total harmonic distortion (THD) and fast current response to obtain high performance. However, the PMSM in a FESS needs to operate at high speed, making it difficult to obtain a low current THD. One of the reasons for the high current THD is that the sampling frequency of the control system is limited. This paper proposes a new multirate model predictive current control (MRMPCC) for the PMSM in a FESS to achieve low current THD and fast current response. The proposed MRMPCC can obtain a high control frequency at a low sampling frequency, which can reduce the current THD. The MRMPCC can also compensate for control delay, which can offer a fast current response to the FESS. In addition, the MRMPCC employs feedforward decoupling to simplify the predictive model. The proposed MRMPCC is implemented on a 36000-rpm magnetic suspension FESS prototype. Compared to a conventional proportional integral differential (PID) control and common dead-beat model predictive current control (DBMPCC), the proposed MRMPCC can provide lower current THD, higher efficiency, and faster current response for a FESS. The proposed MRMPCC is verified experimentally.

Distributed electric propulsion technology has great potential and advantages in the development of drones. In this paper, to study the slipstream effect of distributed propellers, the actuator disk method was used to verify a single propeller, and the calculated thrust was in good agreement with the test results. Then, based on the actuator disk method, the influence of different installation positions on the slipstream effect was studied, and the distributed propeller layout was optimized by a genetic algorithm to improve the low-speed performance of the unmanned aerial vehicle (UAV) during the take-off phase and increase the cruise duration. The analysis results showed that the lift of the wing will be larger when the propellers are higher than the wing. The wing lift and drag of the counter-rotating are less than those of the co-rotating. Compared with the original layout, the lift coefficient of the optimized distributed propeller layout is significantly increased by 30.97%, while the lift/drag ratio is increased by 7.34%. Finally, we designed the test platform and qualitatively verified the calculated results without quantitative verification.

The integrity of a point cloud frequently suffers from discontinuous material surfaces or coarse sensor resolutions. Existing methods focus on reconstructing the overall structure, but salient points or small irregular surfaces are difficult to be predicted. Toward this issue, we propose a new end-to-end neural network for point cloud completion. To avoid non-uniform point density, the regular voxel centers are selected as reference points. The encoder and decoder are designed with Patchify, transformers, and multilayer perceptrons. An implicit classifier is incorporated in the decoder to mark the valid voxels that are allowed for diffusion after removing vacant grids from completion. With newly designed loss function, the classifier is trained to learn the contours, which helps to identify the grids that are difficult to be judged for diffusion. The effectiveness of the proposed model is validated in the experiments on the indoor ShapeNet dataset, the outdoor KITTI dataset, and the airbone laser dataset by competing with state-of-the-art methods, which show that our method can predict more accurate point coordinates with rich details and uniform point distributions.

BackgroundOrexin-A (OXA) is a polypeptide produced in the hypothalamus, which binds to specific receptors and exerts multiple physiological effects. Autophagy plays a vital role in early brain injury (EBI) after intracerebral hemorrhage (ICH). However, the relationship between OXA and autophagy after ICH has not been confirmed.MethodsIn this study, the protective role of OXA was investigated in a model of hemin-induced injury in PC12 cells and blood-injection ICH model in rats, and its potential molecular mechanism was clarified. Neurobehavioral tests, brain water content, and pathologic morphology were assessed after ICH. Cell survival rate was determined using Cell Counting Kit-8 (CCK-8), while apoptosis was detected using flow cytometry. The autophagy protein LC3 that was originally identified as microtubule-associated protein 1 light 3 was evaluated by immunohistochemistry. The ultrastructural changes of cells following ICH were observed by transmission electron microscopy. Western blotting was performed to determine the expression levels of LC3, p62/SQSTM1 (p62), phosphorylated extracellular signal-regulated kinase 1/2 (p-ERK1/2), total extracellular signal-regulated kinase 1/2 (t-ERK1/2), mammalian target of rapamycin (mTOR), and phosphorylated mammalian target of rapamycin (p-mTOR).ResultsOXA treatment significantly improved neurofunctional outcomes, reduced brain edema, and alleviated neuronal apoptosis. OXA administration upregulated p-mTOR and p62, while it downregulated p-ERK1/2 and LC3; this effect was reversed by the orexin receptor 1 (OXR1) antagonist SB-334867.ConclusionsThis study demonstrates that OXA suppresses autophagy via the OXR1-mediated ERK/mTOR signaling pathway to exert neuroprotective effects, and it might provide a novel therapeutic approach in patients suffering from ICH.

The quality and safety of agricultural products is very important for farmers' professional cooperatives. This study incorporates the government, farmers' professional cooperatives and farmers into the evolutionary game model to explore the game relationship and evolutionary path of decision-making among the three parties related to the quality and safety of agricultural products. Through the dynamic analysis of decision-making replication, the analysis of strategy evolution stability and the verification of numerical simulation experiments, it is shown that the decision-making behavior of the government, farmers' professional cooperatives and farmers under the conditions of agricultural product quality and safety is the result of the game between three stakeholders; the government, farmers' professional cooperatives and farmers The evolution process of the decision-making behavior to the ideal state is affected by many factors, and the value ranges of different factors have different effects on the convergence speed of the three stakeholders to the ideal state; when certain conditions are met, the government, farmers' professional cooperatives and farmers' three The decision-making behavior of each stakeholder can evolve into an ideal state, and effective government supervision can promote the cooperatives to manage the green production of farmers, and then effectively encourage farmers to take the initiative in green production. In order to provide useful suggestions for the government to make safety supervision decisions, to effectively manage cooperatives, and to stimulate active green production behaviors by farmers.

Over one hundred spatiotemporal fusion algorithms have been proposed, but convolutional neural networks trained with large amounts of data for spatiotemporal fusion have not shown significant advantages. In addition, no attention has been paid to whether fused images can be used for change detection. These two issues are addressed in this work. A new dataset consisting of nine pairs of images is designed to benchmark the accuracy of neural networks using one-pair spatiotemporal fusion with neural-network-based models. Notably, the size of each image is significantly larger compared to other datasets used to train neural networks. A comprehensive comparison of the radiometric, spectral, and structural losses is made using fourteen fusion algorithms and five datasets to illustrate the differences in the performance of spatiotemporal fusion algorithms with regard to various sensors and image sizes. A change detection experiment is conducted to test if it is feasible to detect changes in specific land covers using the fusion results. The experiment shows that convolutional neural networks can be used for one-pair spatiotemporal fusion if the sizes of individual images are adequately large. It also confirms that the spatiotemporally fused images can be used for change detection in certain scenes.

Mosaicking of remote sensing images stitches images of different moments or sensors to produce a new image under a uniform geographic coordinate system. In a mosaicking process, the critical enblending operation is divided into color balance, seamline finder, and fusion of overlapping areas, which is still challenging to maintain color consistency and data fidelity. In this article, a new mosaicking framework using spatiotemporal fusion is proposed to solve the enblending issue. Two additional low-resolution reference images are introduced for each mosaicking image. With spatiotemporal fusion methods, all mosaicking images are reconstructed to a uniform time, then the combination of overlapping areas become easy. Furthermore, a new spatiotemporal fusion method is proposed by cascading enhanced deep neural networks to fuse images quickly and effectively. In the validation procedure, the proposed method is compared with eight color harmony methods or tools by mosaicking the red, green, and blue bands of Landsat-8 images with images from the moderate-resolution imaging spectroradiometer as the reference. The digital evaluations and visual comparisons demonstrate that the newly method outweighs majority methods regarding to the radiometric, structural, and spectral fidelity, which proves the feasibility of our new enblending method.

The primary factors that impact food safety are the quality and safety of agricultural products. The quality and safety of agricultural products are closely related to farmers’ decision-making behavior. In addition, the decision-making behavior of the government and consumers is the key factor of farmers’ decision-making behavior. The study explores the ideal state among the government, farmers and consumers to promote the improvement of the quality of agricultural products, thereby contributing to food safety. The study includes the government, farmers and consumers in an evolutionary game model to explore game relationships and evolution paths of decision making among the three stakeholders. The dynamic analysis of the decision replication, the stability analysis of strategy evolution, and the numerical simulation experiment verify the following: the evolutionary process of the decision-making behavior of the three stakeholders to an ideal state is affected by many factors, and the value ranges of the different factors have different impacts on the convergence speed of the three stakeholders to an ideal state; When certain conditions are met, the decision-making behaviors of the three stakeholders can evolve to an ideal state, and effective government supervision can promote the transformation of farmers from nongreen to green pesticide application behavior, and encourage consumers to purchase high-quality agricultural products.

The cellular and molecular mechanisms in pathogenesis and development of epilepsy are still unclear. Specific inflammatory mediators and immune cells may play an important role. The aim of the present study was to investigate the temporal and spatial changes of Th17, Tregs, and related cytokines in epilepsy lesions. LiCl-pilocarpine-induced temporal lobe epilepsy (TLE) rat models were established, sensorimotor function was examined using modified neurological severity score (mNSS), cognitive function was evaluated by Morris water maze (MWM) test, pathological damages were detected by H&E staining and Nissl staining, helper T cells 17 (Th17), regulatory CD4+ T cells (Tregs), and their related cytokines were detected by Western blotting and immune staining. Results showed that Th17 and its related cytokines in epilepsy lesions played a role mainly at acute phase of epilepsy, and they were positively correlated with the pathological changes in the hippocampus and neurological and cognitive dysfunction caused by epilepsy. Conversely, Tregs and their related cytokines mainly played a role at progressive phase and had the opposite effect. Th17 and Tregs restricted each other during the recovery phase to achieve functional balance. Our results suggested that Th17, Tregs, and related cytokines in epilepsy lesions played an important role in the pathogenesis and development of epilepsy and balancing Th17 and Tregs may be efficacious therapeutics for patients with epilepsy.

The composite material flywheel rotor of a flywheel energy storage system (FESS) has a low natural frequency. When the system suffers from noise interference, the magnetic bearing generates a force with the same frequency as the natural frequency and causes vibration to occur. Thus, it is necessary to suppress the natural vibration of the magnetic suspended (MS) FESS. The LMS adaptive notch filter is generally adopted for vibration suppression. The vibration suppression performance of the system is different when the insertion position of the notch filter is different. This paper analyzes the influence of the notch filter in different insertion positions of the control system. Through the transfer function from noise to magnetic bearing force, theoretical analysis of the influence of different positions of the notch filter is performed. Corresponding experiments are performed in a 500 kW MS FESS prototype. The theoretical analysis is verified experimentally.

The spatial, sensor, and temporal differences can be observed in the process of spatiotemporal fusion because source images are from different sensors or moments. The existing spatiotemporal fusion methods have modelled the temporal difference, but they did not solve the spatial difference or the sensor difference to build complete models. In this paper, a step-by-step modelling framework is proposed, and three models are designed based on deep neural networks to model the spatial difference, sensor difference, and temporal difference in a separate and explicit way. The spatial difference is modelled with cascaded dual regression networks. The sensor difference is simulated with a four-layer convolutional neural network. The temporal difference is predicted with a generative adversarial network. The proposed method is compared with six algorithms for the reconstruction of Landsat-7 and Landsat-5 which validates the effectiveness of the spatial fusion strategy. The digital evaluation on radiometric, structural, and spectral loss illustrates that the proposed method can give the optimal performance steadily. The necessity of complete modelling is also tested by connecting the spatial and sensor models of the proposed method with one-pair fusion methods, and the steadily improved performance shows that all the difference models contribute to performance improvement.

Spatiotemporal fusion has got enough attention and many algorithms have been proposed, but its practical stability has not been emphasized yet. Observing that the strategies harnessed by different types of algorithms may lead to various tendencies, an integration strategy is introduced to make full use of the complementarity between different types of spatiotemporal fusion algorithms for better fusion stability. In our method, the images fused by two different types of methods are decomposed into components denoting strength, structure, and mean intensity, which are combined separately involving a characteristic analysis. The proposed method is compared with seven algorithms of four types by reconstructing Landsat-8, Landsat-7, and Landsat-5 images to validate the effectiveness of the spatial fusion strategy. The digital evaluation on radiometric, structural, and spectral loss illustrates that the proposed method can reach or approach the optimal performance steadily.

Meteorological satellites are usually operated at high temporal resolutions, but the spatial resolutions are too poor to identify ground content. Super-resolution is an economic way to enhance spatial details, but the feasibility is not validated for meteorological images due to the absence of benchmarking data. In this work, we propose the FY4ASRgray and FY4ASRcolor datasets to assess super-resolution algorithms on meteorological applications. The features of cloud sensitivity and temporal continuity are linked to the proposed datasets. To test the usability of the new datasets, five state-of-the-art super-resolution algorithms are gathered for contest. Shift learning is used to shorten the training time and improve the parameters. Methods are modified to deal with the 16-bit challenge. The reconstruction results are demonstrated and evaluated regarding the radiometric, structural, and spectral loss, which gives the baseline performance for detail enhancement of the FY4A satellite images. Additional experiments are made on FY4ASRcolor for sequence super-resolution, spatiotemporal fusion, and generalization test for further performance test.

A series of exohedral actinide borospherenes, An&Bm, and endohedral borospherenes, An@Bn (An=U, Np, Pu; m = 28, 32, 34, 36, 38, 40; n = 36, 38, 40), have been characterized by density functional theory calculations. The electronic structures, chemical bond topological properties and spectra have been systematically investigated. It was found that An@Bn is more stable than An&Bn in terms of structure and energy, and UB36 in an aqueous solution is the most stable molecular in this research. The IR and UV-vis spectra of An&Bm and An@Bn are computationally predicted to facilitate further experimental investigations. Charge-transfer spectroscopy decomposes the total UV-Vis absorption curve into the contributions of different excitation features, allowing insight into what form of electronic excitation the UV–Vis absorption peak is from the perspective of charge transfer between the An atoms and borospherenes.

A space vector pulse width modulation (SVPWM) algorithm is an important part of the permanent magnet synchronous machine (PMSM) drive to achieve direct current (DC) to alternating current (AC) conversion. The execution of the conventional SVPWM algorithm is a complex process which will limit the sampling frequency of the high-speed PMSM drive. Low sampling frequency will cause high current total harmonic distortion (THD) and eddy current loss. To increase the sampling frequency, this paper proposes a novel simplified SVPWM algorithm. The proposed SVPWM algorithm turns the vector composition problem of the conventional SVPWM algorithm into an optimization problem of the dwell time of the basic vector. The proposed SVPWM algorithm has an optimal vector dwell time (OVDT). The dwell time of the basic vector can be directly calculated by solving the optimization problem. The proposed SVPWM algorithm does not need sector identification compared to the conventional algorithm. The experiments of the proposed SVPWM algorithm are performed in a high-speed PMSM drive of a flywheel energy storage system (FESS). Compared to the conventional SVPWM algorithm, the execution time of the proposed SVPWM algorithm is reduced by 38%. By using the proposed SVPWM algorithm, the sampling frequency can be increased from 33 kHz to 40 kHz. With the higher sampling frequency, the current THD is reduced by 25.6%. The effectiveness of the proposed simplified SVPWM algorithm is verified experimentally.

To realize the hyperstatic performance index of a magnetically suspended flywheel and simultaneously suppress the vibration force caused by mass imbalance and sensor runout, a compound control method based on a repetitive controller and displacement force compensation of the synchronous force is proposed. First, the mechanism of different interference vibration forces is analyzed by establishing a model of the magnetically suspended flywheel. The analysis shows that the x–y direction is symmetric, and the flywheel structure has symmetry. Second, considering the symmetry of the x- and y-directions, the x-direction is taken as an example for analysis, the parameter design and stability analysis are carried out, and the range of parameters of the compound repetitive control method is obtained. Finally, a flywheel with different speeds is simulated. It was found that the vibration force of each frequency can be suppressed by the compound control method, and the inhibition rate of the vibration force can reach as much as 95%. The results show that the unbalanced vibration and vibration force caused by the sensor runout can be effectively suppressed by using the compound repetitive control method.

The fast and accurate prediction of crop yield at the regional scale is of great significance to food policies or trade. In this study, a new model is developed to predict the yield of oilseed rape from high-resolution remote sensing images. In order to derive this model, the ground experiment and remote sensing data analysis are carried out successively. In the ground experiment, the leaf area index (LAI) of four growing stages are measured, and a regression model is established to predict yield from ground LAI. In the remote sensing analysis, a new model is built to predict ground LAI from Gaofen-1 images where the simple ratio vegetation index at the bolting stage and the VARIgreen vegetation index at the flowering stage are used. The WOFOSTWOrld FOod STudy (WOFOST) crop model is used to generate time-series ground LAI from discontinuous ground LAI, which is calibrated coarsely with the MODerate resolution imaging spectroradiometer LAI product and finely with the ground-measured data. By combining the two conclusive formulas, an estimation model is built from Gaofen-1 images to the yield of oilseed rape. The effectiveness of the proposed model is verified in Wuxue City, Hubei Province from 2014 to 2019, with the pyramid bottleneck residual network to extract oilseed rape planting areas, the proposed model to estimate yields, and the China statistical yearbooks for comparison. The validation shows that the prediction error of the proposed algorithm is less than 5.5%, which highlights the feasibility of our method for accurate prediction of the oilseed rape yield in a large area.

Mechanical resonance occurs during the operation of a maglev inertially stabilized platform (MISP). The MISP is driven by the motor gear and the gear clearance changes during commutation, which makes the resonance frequency variable. Conventional notch filters with constant parameters are not ideal for variable frequency resonance suppression. In this study, the forgetting factor in the recursive least squares (RLS) algorithm was improved to estimate the resonance frequency online. By adjusting the frequency parameters of the notch filter, the proposed measures can effectively suppress the variable frequency resonance in a dual-stage MISP system. Finally, a simulation example is presented to validate the arguments in this study.

The visualization of near infrared hyperspectral images is valuable for quick view and information survey, whereas methods using band selection or dimension reduction fail to produce good colors as reasonable as corresponding multispectral images. In this paper, an end-to-end neural network of hyperspectral visualization is proposed, based on the convolutional neural networks, to transform a hyperspectral image of hundreds of near infrared bands to a three-band image. Supervised learning is used to train the network where multispectral images are targeted to reconstruct naturally looking images. Each pair of the training images shares the same geographic location and similar moments. The generative adversarial framework is used with an adversarial network to improve the training of the generating network. In the experimental procedure, the proposed method is tested for the near infrared bands of EO-1 Hyperion images with LandSat-8 images as the benchmark, which is compared with five state-of-the-art visualization algorithms. The experimental results show that the proposed method performs better in producing naturally looking details and colors for near infrared hyperspectral images.

This paper investigates energy harvesting based multiuser system with large-scale distributed antennas, where a large number of remote antenna units (RAUs) are evenly separated across multiple circles. An efficient wireless energy and information transmission protocol is proposed. To save the signaling and the radio frequency chains overheads, the RAU with the shortest distance towards a user equipment (UE) is employed for the downlink wireless energy transfer (WET). In the uplink phase, we analyze the probability of wireless information transmission (WIT) of UEs. Then, linear zero-forcing detection and minimum-mean-square error are used to separate the data information among all the UEs that satisfy the requirement of WIT. The asymptotic throughput for an arbitrary UE is derived in closed-form. The time fraction used for the WET is optimized through maximizing the asymptotic throughput. Numerical and simulation results are given to verify the theoretical analysis, and bring to light the time fraction planning and the RAUs deployment for the system.

Reconstructing images from their noisy and incomplete measurements is always a challenge especially for medical MR image with important details and features. This work proposes a novel dictionary learning model that integrates two sparse regularization methods: the total generalized variation (TGV) approach and adaptive dictionary learning (DL). In the proposed method, the TGV selectively regularizes different image regions at different levels to avoid oil painting artifacts largely. At the same time, the dictionary learning adaptively represents the image features sparsely and effectively recovers details of images. The proposed model is solved by variable splitting technique and the alternating direction method of multiplier. Extensive simulation experimental results demonstrate that the proposed method consistently recovers MR images efficiently and outperforms the current state-of-the-art approaches in terms of higher PSNR and lower HFEN values.

JNI in the Android platform is often observed with low efficiency and high coding complexity. Although many researchers have investigated the JNI mechanism, few of them solve the efficiency and the complexity problems of JNI in the Android platform simultaneously. In this paper, a hybrid polylingual object (HPO) model is proposed to allow a CAR object being accessed as a Java object and as vice in the Dalvik virtual machine. It is an acceptable substitute for JNI to reuse the CAR-compliant components in Android applications in a seamless and efficient way. The metadata injection mechanism is designed to support the automatic mapping and reflection between CAR objects and Java objects. A prototype virtual machine, called HPO-Dalvik, is implemented by extending the Dalvik virtual machine to support the HPO model. Lifespan management, garbage collection, and data type transformation of HPO objects are also handled in the HPO-Dalvik virtual machine automatically. The experimental result shows that the HPO model outweighs the standard JNI in lower overhead on native side, better executing performance with no JNI bridging code being demanded.

Fusion of remote sensing images with different spatial and temporal resolutions is highly needed by diverse earth observation applications. A small number of spatiotemporal fusion methods using sparse representation appear to be more promising than traditional linear mixture methods in reflecting abruptly changing terrestrial content. However, one of the main difficulties is that the results of sparse representation have reduced expressional accuracy; this is due in part to insufficient prior knowledge. For remote sensing images, the cluster and joint structural sparsity of the sparse coefficients could be employed as a priori knowledge. In this paper, a new optimization model is constructed with the semi-coupled dictionary learning and structural sparsity to predict the unknown high-resolution image from known images. Specifically, the intra-block correlation and cluster-structured sparsity are considered for single-channel reconstruction, and the inter-band similarity of joint-structured sparsity is considered for multichannel reconstruction, and both are implemented with block sparse Bayesian learning. The detailed optimization steps are given iteratively. In the experimental procedure, the red, green, and near-infrared bands of Landsat-7 and Moderate Resolution Imaging Spectrometer (MODIS) satellites are put to fusion with root mean square errors to check the prediction accuracy. It can be concluded from the experiment that the proposed methods can produce higher quality than state-of-the-art methods.

Because of the contradiction between the spatial and temporal resolution of remote sensing images (RSI) and quality loss in the process of acquisition, it is of great significance to reconstruct RSI in remote sensing applications. Recent studies have demonstrated that reference image-based reconstruction methods have great potential for higher reconstruction performance, while lacking accuracy and quality of reconstruction. For this application, a new compressed sensing objective function incorporating a reference image as prior information is developed. We resort to the reference prior information inherent in interior and exterior data simultaneously to build a new generalized nonconvex low-rank approximation framework for RSI reconstruction. Specifically, the innovation of this paper consists of the following three respects: (1) we propose a nonconvex low-rank approximation for reconstructing RSI; (2) we inject reference prior information to overcome over smoothed edges and texture detail losses; (3) on this basis, we combine conjugate gradient algorithms and a single-value threshold (SVT) simultaneously to solve the proposed algorithm. The performance of the algorithm is evaluated both qualitatively and quantitatively. Experimental results demonstrate that the proposed algorithm improves several dBs in terms of peak signal to noise ratio (PSNR) and preserves image details significantly compared to most of the current approaches without reference images as priors. In addition, the generalized nonconvex low-rank approximation of our approach is naturally robust to noise, and therefore, the proposed algorithm can handle low resolution with noisy inputs in a more unified framework.