A method to remove stripes from remote sensing images is proposed based on statistics and a new image enhancement method.The overall processing steps for improving the quality of remote sensing images are introduced t...A method to remove stripes from remote sensing images is proposed based on statistics and a new image enhancement method.The overall processing steps for improving the quality of remote sensing images are introduced to provide a general baseline.Due to the differences in satellite sensors when producing images,subtle but inherent stripes can appear at the stitching positions between the sensors.These stitchingstripes cannot be eliminated by conventional relative radiometric calibration.The inherent stitching stripes cause difficulties in downstream tasks such as the segmentation,classification and interpretation of remote sensing images.Therefore,a method to remove the stripes based on statistics and a new image enhancement approach are proposed in this paper.First,the inconsistency in grayscales around stripes is eliminated with the statistical method.Second,the pixels within stripes are weighted and averaged based on updated pixel values to enhance the uniformity of the overall image radiation quality.Finally,the details of the images are highlighted by a new image enhancement method,which makes the whole image clearer.Comprehensive experiments are performed,and the results indicate that the proposed method outperforms the baseline approach in terms of visual quality and radiation correction accuracy.展开更多
With the adoption of cutting-edge communication technologies such as 5G/6G systems and the extensive development of devices,crowdsensing systems in the Internet of Things(IoT)are now conducting complicated video analy...With the adoption of cutting-edge communication technologies such as 5G/6G systems and the extensive development of devices,crowdsensing systems in the Internet of Things(IoT)are now conducting complicated video analysis tasks such as behaviour recognition.These applications have dramatically increased the diversity of IoT systems.Specifically,behaviour recognition in videos usually requires a combinatorial analysis of the spatial information about objects and information about their dynamic actions in the temporal dimension.Behaviour recognition may even rely more on the modeling of temporal information containing short-range and long-range motions,in contrast to computer vision tasks involving images that focus on understanding spatial information.However,current solutions fail to jointly and comprehensively analyse short-range motions between adjacent frames and long-range temporal aggregations at large scales in videos.In this paper,we propose a novel behaviour recognition method based on the integration of multigranular(IMG)motion features,which can provide support for deploying video analysis in multimedia IoT crowdsensing systems.In particular,we achieve reliable motion information modeling by integrating a channel attention-based short-term motion feature enhancement module(CSEM)and a cascaded long-term motion feature integration module(CLIM).We evaluate our model on several action recognition benchmarks,such as HMDB51,Something-Something and UCF101.The experimental results demonstrate that our approach outperforms the previous state-of-the-art methods,which confirms its effective-ness and efficiency.展开更多
Due to hardware limitations,existing hyperspectral(HS)camera often suffer from low spatial/temporal resolution.Recently,it has been prevalent to super-resolve a low reso-lution(LR)HS image into a high resolution(HR)HS...Due to hardware limitations,existing hyperspectral(HS)camera often suffer from low spatial/temporal resolution.Recently,it has been prevalent to super-resolve a low reso-lution(LR)HS image into a high resolution(HR)HS image with a HR RGB(or mul-tispectral)image guidance.Previous approaches for this guided super-resolution task often model the intrinsic characteristic of the desired HR HS image using hand-crafted priors.Recently,researchers pay more attention to deep learning methods with direct supervised or unsupervised learning,which exploit deep prior only from training dataset or testing data.In this article,an efficient convolutional neural network-based method is presented to progressively super-resolve HS image with RGB image guidance.Specif-ically,a progressive HS image super-resolution network is proposed,which progressively super-resolve the LR HS image with pixel shuffled HR RGB image guidance.Then,the super-resolution network is progressively trained with supervised pre-training and un-supervised adaption,where supervised pre-training learns the general prior on training data and unsupervised adaptation generalises the general prior to specific prior for variant testing scenes.The proposed method can effectively exploit prior from training dataset and testing HS and RGB images with spectral-spatial constraint.It has a good general-isation capability,especially for blind HS image super-resolution.Comprehensive experimental results show that the proposed deep progressive learning method out-performs the existing state-of-the-art methods for HS image super-resolution in non-blind and blind cases.展开更多
Point-of-interest(POI) recommendation is a popular topic on location-based social networks(LBSNs).Geographical proximity,known as a unique feature of LBSNs,significantly affects user check-in behavior.However,most of ...Point-of-interest(POI) recommendation is a popular topic on location-based social networks(LBSNs).Geographical proximity,known as a unique feature of LBSNs,significantly affects user check-in behavior.However,most of prior studies characterize the geographical influence based on a universal or personalized distribution of geographic distance,leading to unsatisfactory recommendation results.In this paper,the personalized geographical influence in a two-dimensional geographical space is modeled using the data field method,and we propose a semi-supervised probabilistic model based on a factor graph model to integrate different factors such as the geographical influence.Moreover,a distributed learning algorithm is used to scale up our method to large-scale data sets.Experimental results based on the data sets from Foursquare and Gowalla show that our method outperforms other competing POI recommendation techniques.展开更多
The reconstruction control of modular self-reconfigurable spacecraft (MSRS) is addressed using an adaptive sliding mode control (ASMC) scheme based on time-delay estimation (TDE) technology. In contrast to the ground,...The reconstruction control of modular self-reconfigurable spacecraft (MSRS) is addressed using an adaptive sliding mode control (ASMC) scheme based on time-delay estimation (TDE) technology. In contrast to the ground, the base of the MSRS is floating when assembled in orbit, resulting in a strong dynamic coupling effect. A TED-based ASMC technique with exponential reaching law is designed to achieve high-precision coordinated control between the spacecraft base and the robotic arm. TDE technology is used by the controller to compensate for coupling terms and uncertainties, while ASMC can augment and improve TDE’s robustness. To suppress TDE errors and eliminate chattering, a new adaptive law is created to modify gain parameters online, ensuring quick dynamic response and high tracking accuracy. The Lyapunov approach shows that the tracking errors are uniformly ultimately bounded (UUB). Finally, the on-orbit assembly process of MSRS is simulated to validate the efficacy of the proposed control scheme. The simulation results show that the proposed control method can accurately complete the target module’s on-orbit assembly, with minimal perturbations to the spacecraft’s attitude. Meanwhile, it has a high level of robustness and can effectively eliminate chattering.展开更多
The dynamic characteristics related to micro-motions, such as mechanical vibration or rotation, play an essential role in classifying and recognizing ballistic targets in the midcourse, and recent researches explore w...The dynamic characteristics related to micro-motions, such as mechanical vibration or rotation, play an essential role in classifying and recognizing ballistic targets in the midcourse, and recent researches explore ways of extracting the micro-motion features from radar signals of ballistic targets. In this paper, we focus on how to investigate the micro-motion dynamic characteristics of the ballistic targets from the signals based on infrared (IR) detection, which is mainly achieved by analyzing the periodic fluctuation characteristics of the target IR irradiance intensity signatures. Simulation experiments demonstrate that the periodic characteristics of IR signatures can be used to distinguish different micro motion types and estimate related parameters. Consequently, this is possible to determine the micro-motion dynamics of ballistic targets based on IR detection.展开更多
Most of the current object detection algorithms use pretrained models that are trained on ImageNet and then fine-tuned in the network,which can achieve good performance in terms of general object detectors.However,in ...Most of the current object detection algorithms use pretrained models that are trained on ImageNet and then fine-tuned in the network,which can achieve good performance in terms of general object detectors.However,in the field of remote sensing image object detection,as pretrained models are significantly different from remote sensing data,it is meaningful to explore a train-fromscratch technique for remote sensing images.This paper proposes an object detection framework trained from scratch,SRS-Net,and describes the design of a densely connected backbone network to provide integrated hidden layer supervision for the convolution module.Then,two necessary improvement principles are proposed:studying the role of normalization in the network structure,and improving data augmentation methods for remote sensing images.To evaluate the proposed framework,we performed many ablation experiments on the DIOR,DOTA,and AS datasets.The results show that whether using the improved backbone network,the normalization method or training data enhancement strategy,the performance of the object detection network trained from scratch increased.These principles compensate for the lack of pretrained models.Furthermore,we found that SRS-Net could achieve similar to or slightly better performance than baseline methods,and surpassed most advanced general detectors.展开更多
Object detection in Remote Sensing(RS)has achieved tremendous advances in recent years,but it remains challenging for rotated object detection due to cluttered backgrounds,dense object arrangements and the wide range ...Object detection in Remote Sensing(RS)has achieved tremendous advances in recent years,but it remains challenging for rotated object detection due to cluttered backgrounds,dense object arrangements and the wide range of size variations among objects.To tackle this problem,Dense Context Feature Pyramid Network(DCFPN)and a powerα-Gaussian loss are designed for rotated object detection in this paper.The proposed DCFPN can extract multi-scale information densely and accurately by leveraging a dense multi-path dilation layer to cover all sizes of objects in remote sensing scenarios.For more accurate detection while avoiding bottlenecks such as boundary discontinuity in rotated bounding box regression,a-Gaussian loss,a unified power generalization of existing Gaussian modeling losses is proposed.Furthermore,the properties ofα-Gaussian loss are analyzed comprehensively for a wider range of applications.Experimental results on four datasets(UCAS-AOD,HRSC2016,DIOR-R,and DOTA)show the effectiveness of the proposed method using different detectors,and are superior to the existing methods in both feature extraction and bounding box regression。展开更多
In order to rapidly and accurately detect infrared small and dim targets in the infrared image of complex scene collected by virtual prototyping of space-based downward-looking multiband detection,an improved detectio...In order to rapidly and accurately detect infrared small and dim targets in the infrared image of complex scene collected by virtual prototyping of space-based downward-looking multiband detection,an improved detection algorithm of infrared small and dim target is proposed in this paper.Firstly,the original infrared images are changed into a new infrared patch tensor mode through data reconstruction.Then,the infrared small and dim target detection problems are converted to low-rank tensor recovery problems based on tensor nuclear norm in accordance with patch tensor characteristics,and inverse variance weighted entropy is defined for self-adaptive adjustment of sparseness.Finally,the low-rank tensor recovery problem with noise is solved by alternating the direction method to obtain the sparse target image,and the final small target is worked out by a simple partitioning algorithm.The test results in various spacebased downward-looking complex scenes show that such method can restrain complex background well by virtue of rapid arithmetic speed with high detection probability and low false alarm rate.It is a kind of infrared small and dim target detection method with good performance.展开更多
The electrical performance of radomes on high-speed aircraft can be influenced by the thermal and mechanical loads produced during high-speed flight,which can affect the detection dis-tance and accuracy of the guidanc...The electrical performance of radomes on high-speed aircraft can be influenced by the thermal and mechanical loads produced during high-speed flight,which can affect the detection dis-tance and accuracy of the guidance system.This paper presents a new method that uses the Finite Difference Time Domain(FDTD)method to calculate the electrical performance of radomes under Thermo-Mechanical-Electrical(TME)coupling.This method can accurately characterize the effects of material dielectric temperature drift and structural deformation on the electrical performance of the radome under flight conditions,enabling high-precision full-wave calculations of the broadband electrical performance of the radome.The method initiates by utilizing a Finite Element Grid Model(FE-GM)of the radome to sequentially acquire the radome's response temperature field and structural deformation field through thermal and mechanical simulations.Subsequently,spatial mapping techniques are developed to accurately incorporate the dielectric temperature drift and structural deformation of the radome into its Yee grid Electromagnetic(EM)simulation model.A verification case was designed to test the proposed method,and the results confirmed its high computational accuracy.Additionally,the effectiveness and necessity of the method were further demonstrated by analyzing the electrical performance of a fused silica ceramic radome used on a high-speed aircraft.展开更多
The earth observation satellites(EOSs)scheduling problem for emergency tasks often presents many challenges.For example,the scheduling calculation should be completed in seconds,the scheduled task rate is supposed to ...The earth observation satellites(EOSs)scheduling problem for emergency tasks often presents many challenges.For example,the scheduling calculation should be completed in seconds,the scheduled task rate is supposed to be as high as possible,the disturbance measure of the scheme should be as low as possible,which may lead to the loss of important observation opportunities and data transmission delays.Existing scheduling algorithms are not designed for these requirements.Consequently,we propose a rolling horizon strategy(RHS)based on event triggering as well as a heuristic algorithm based on direct insertion,shifting,backtracking,deletion,and reinsertion(ISBDR).In the RHS,the driven scheduling mode based on the emergency task arrival and control station time window events are designed to transform the long-term,large-scale problem into a short-term,small-scale problem,which can improve the schedulability of the original scheduling scheme and emergency response sensitivity.In the ISBDR algorithm,the shifting rule with breadth search capability and backtracking rule with depth search capability are established to realize the rapid adjustment of the original plan and improve the overall benefit of the plan and early completion of emergency tasks.Simultaneously,two heuristic factors,namely the emergency task urgency degree and task conflict degree,are constructed to improve the emergency task scheduling guidance and algorithm efficiency.Finally,we conduct extensive experiments by means of simulations to compare the algorithms based on ISBDR and direct insertion,shifting,deletion,and reinsertion(ISDR).The results demonstrate that the proposed algorithm can improve the timeliness of emergency tasks and scheduling performance,and decrease the disturbance measure of the scheme,therefore,it is more suitable for emergency task scheduling.展开更多
For the detection of marine ship objects in radar images, large-scale networks based on deep learning are difficult to be deployed on existing radar-equipped devices. This paper proposes a lightweight convolutional ne...For the detection of marine ship objects in radar images, large-scale networks based on deep learning are difficult to be deployed on existing radar-equipped devices. This paper proposes a lightweight convolutional neural network, LiraNet, which combines the idea of dense connections, residual connections and group convolution, including stem blocks and extractor modules.The designed stem block uses a series of small convolutions to extract the input image features, and the extractor network adopts the designed two-way dense connection module, which further reduces the network operation complexity. Mounting LiraNet on the object detection framework Darknet, this paper proposes Lira-you only look once(Lira-YOLO), a lightweight model for ship detection in radar images, which can easily be deployed on the mobile devices. Lira-YOLO's prediction module uses a two-layer YOLO prediction layer and adds a residual module for better feature delivery. At the same time, in order to fully verify the performance of the model, mini-RD, a lightweight distance Doppler domain radar images dataset, is constructed. Experiments show that the network complexity of Lira-YOLO is low, being only 2.980 Bflops, and the parameter quantity is smaller, which is only 4.3 MB. The mean average precision(mAP) indicators on the mini-RD and SAR ship detection dataset(SSDD) reach 83.21% and 85.46%, respectively,which is comparable to the tiny-YOLOv3. Lira-YOLO has achieved a good detection accuracy with less memory and computational cost.展开更多
In the high speed target environment,there exists serious Doppler effect in the low pulse repetition frequency（LPRF） modulated frequency stepped frequency（MFSF） radar signal.The velocity range of the target is lar...In the high speed target environment,there exists serious Doppler effect in the low pulse repetition frequency（LPRF） modulated frequency stepped frequency（MFSF） radar signal.The velocity range of the target is large and the velocity is high ambiguous,so the single method is difficult to satisfy the velocity measurement requirement.For this problem,a novel method is presented,it is a combination of cross-correlation inner frame velocity measurement and range-Doppler coupling velocity measurement.The cross-correlation inner frame method,overcoming the low Doppler tolerance of the cross-correlation between frames,can obtain the coarse velocity of the high speed target,and then the precision velocity can be obtained with the range-Doppler coupling method.The simulation results confirm the method is effective,and also it is well real-time and easy to the project application.展开更多
Satellite networking communications in navigation satellite system and spacebased deep space exploration have the features of a long delay and high bit error rate （BER）. Through analyzing the advantages and disadvan...Satellite networking communications in navigation satellite system and spacebased deep space exploration have the features of a long delay and high bit error rate （BER）. Through analyzing the advantages and disadvantages of the Consulta tive Committee for the Space Data System （CCSDS） file delivery protocol （CFDP）, a new improved repeated sending file delivery protocol （RSFDP） based on the adaptive repeated sending is put forward to build an efficient and reliable file transmission. According to the estimation of the BER of the transmission link, RSFDP repeatedly sends the lost protocol data units （PDUs） at the stage of the retransmission to improve the success rate and reduce time of the retransmission. Theoretical analyses and results of the Opnet simulation indicate that the performance of RSFDP has significant improvement gains over CFDP in the link with a long delay and high BER. The realizing results based on the space borne filed programmable gate array （FPGA） platform show the applicability of the proposed algorithm.展开更多
Due to the attractive potential in avoiding the elaborate definition of anchor attributes,anchor-free-based deep learning approaches are promising for object detection in remote sensing imagery.Corner Net is one of th...Due to the attractive potential in avoiding the elaborate definition of anchor attributes,anchor-free-based deep learning approaches are promising for object detection in remote sensing imagery.Corner Net is one of the most representative methods in anchor-free-based deep learning approaches.However,it can be observed distinctly from the visual inspection that the Corner Net is limited in grouping keypoints,which significantly impacts the detection performance.To address the above problem,a novel and effective approach,called Group Net,is presented in this paper,which adaptively groups corner specific to the objects based on corner embedding vector and corner grouping network.Compared with the Corner Net,the proposed approach is more effective in learning the semantic relationship between corners and improving remarkably the detection performance.On NWPU dataset,experiments demonstrate that our Group Net not only outperforms the Corner Net with an AP of 12.8%,but also achieves comparable performance to considerable approaches with 83.4%AP.展开更多
Source localization plays an indispensable role in many applications.This paper addresses the directional source localization problem in a three-dimensional(3D)wireless sensor network using hybrid received-signal-stre...Source localization plays an indispensable role in many applications.This paper addresses the directional source localization problem in a three-dimensional(3D)wireless sensor network using hybrid received-signal-strength(RSS)and angle-of-arrival(AOA)measurements.Both the position and transmission orientation of the source are to be estimated.In the considered positioning scenario,the angle and range measurements are respectively corresponding to the AOA model and RSS model that integrates the Gaussian-shaped radiation pattern.Given that the localization problem is non-convex and the unknown parameters therein are coupled together,this paper adopts the second-order cone relaxation and alternating optimization techniques in the proposed estimation algorithm.Moreover,to provide a performance benchmark for any localization method,the corresponding Cramer-Rao lower bounds(CRLB)of estimating the unknown position and transmission orientation of the source are derived.Numerical and simulation results demonstrate that the presented algorithm effectively resolves the problem,and its estimation performance is close to the CRLB for the localization with the hybrid measurements.展开更多
Considering the sparsity of hyperspectral images(HSIs),dictionary learning frameworks have been widely used in the field of unsupervised spectral unmixing.However,it is worth mentioning here that existing dictionary l...Considering the sparsity of hyperspectral images(HSIs),dictionary learning frameworks have been widely used in the field of unsupervised spectral unmixing.However,it is worth mentioning here that existing dictionary learning method-based unmixing methods are found to be short of robustness in noisy contexts.To improve the performance,this study specifically puts forward a new unsupervised spectral unmixing solution.For the reason that the solution only functions in a condition that both endmembers and the abundances meet non-negative con-straints,a model is built to solve the unsupervised spectral un-mixing problem on the account of the dictionary learning me-thod.To raise the screening accuracy of final members,a new form of the target function is introduced into dictionary learning practice,which is conducive to the growing robustness of noisy HSI statistics.Then,by introducing the total variation(TV)terms into the proposed spectral unmixing based on robust nonnega-tive dictionary learning(RNDLSU),the context information under HSI space is to be cited as prior knowledge to compute the abundances when performing sparse unmixing operations.Ac-cording to the final results of the experiment,this method makes favorable performance under varying noise conditions,which is especially true under low signal to noise conditions.展开更多
A new long term integration algorithm is proposed forhigh-dynamic targets, which can resolve the problems of spectrumspread, frequency walk and pseudorandom noise （PRN）code phase curvature caused by the motion of ta...A new long term integration algorithm is proposed forhigh-dynamic targets, which can resolve the problems of spectrumspread, frequency walk and pseudorandom noise （PRN）code phase curvature caused by the motion of targets. This algorithmfirst applies a keystone transform based improved discretepolynomial-phase transform （KT-IDPT） to estimate the Dopplerchirp rate. Then, based on the estimated Doppler chirp rate,dechirping and envelope translation are performed on the partialcorrelation results to correct the spectrum spread and the codephase curvature. The simulation results demonstrate that the proposedmethod has low integration loss and computational burden.展开更多
Generally,Doppler fuze can only estimate actuation delay-time with a limited precision. As an improvement,imaging fuze can estimate actuation delay-time more precisely with the available two-dimensional image of the t...Generally,Doppler fuze can only estimate actuation delay-time with a limited precision. As an improvement,imaging fuze can estimate actuation delay-time more precisely with the available two-dimensional image of the target. In this paper,imprecision of actuation delay-time estimation with Doppler fuze is first analyzed theoretically in brief. Secondly,feasibility analysis and theoretical model of imaging fuze are described,in which a criterion is established for the actuation delay-time based on the image,and then an image based gray-value weighted least square（ GWLS） algorithm is presented to calculate actuation delay-time of the imaging fuze. Finally,a simulation model of missiletarget near-field encounter is established. Simulation results indicate that actuation delay-time of the imaging fuze is estimated more precisely than by the Doppler fuze.展开更多
Laser communication is essential part of maritime-terrestrial-air intelligent communication/sensor network. Among them, different modulation formats would play a unique role in specific applications. Based on Rytov th...Laser communication is essential part of maritime-terrestrial-air intelligent communication/sensor network. Among them, different modulation formats would play a unique role in specific applications. Based on Rytov theory, we discussed system performance of the maritime laser communication with repeated coding technology in several modulation schemes. The closed-form expression of average bit error rate(BER) from weak to moderate atmospheric turbulence described by log-normal distribution is given. Differential phase shift keying(DPSK) modulation, as a potential solution for future maritime laser communication, has attracted a lot of attention. We analyzed the effects of atmospheric turbulence parameters(visibility, refractive index structure coefficient, non-Kolmogorov spectral power-law exponent, turbulence inner scale) and DPSK system parameters(receiver aperture diameter, repeat time) on average BER in detail. Compared with the aperture-averaging effects, the system BER can be well suppressed through increasing repeat time. This work is anticipated to provide a theoretical reference for maritime laser communication systems.展开更多
文摘A method to remove stripes from remote sensing images is proposed based on statistics and a new image enhancement method.The overall processing steps for improving the quality of remote sensing images are introduced to provide a general baseline.Due to the differences in satellite sensors when producing images,subtle but inherent stripes can appear at the stitching positions between the sensors.These stitchingstripes cannot be eliminated by conventional relative radiometric calibration.The inherent stitching stripes cause difficulties in downstream tasks such as the segmentation,classification and interpretation of remote sensing images.Therefore,a method to remove the stripes based on statistics and a new image enhancement approach are proposed in this paper.First,the inconsistency in grayscales around stripes is eliminated with the statistical method.Second,the pixels within stripes are weighted and averaged based on updated pixel values to enhance the uniformity of the overall image radiation quality.Finally,the details of the images are highlighted by a new image enhancement method,which makes the whole image clearer.Comprehensive experiments are performed,and the results indicate that the proposed method outperforms the baseline approach in terms of visual quality and radiation correction accuracy.
基金supported by National Natural Science Foundation of China under grant No.62271125,No.62273071Sichuan Science and Technology Program(No.2022YFG0038,No.2021YFG0018)+1 种基金by Xinjiang Science and Technology Program(No.2022273061)by the Fundamental Research Funds for the Central Universities(No.ZYGX2020ZB034,No.ZYGX2021J019).
文摘With the adoption of cutting-edge communication technologies such as 5G/6G systems and the extensive development of devices,crowdsensing systems in the Internet of Things(IoT)are now conducting complicated video analysis tasks such as behaviour recognition.These applications have dramatically increased the diversity of IoT systems.Specifically,behaviour recognition in videos usually requires a combinatorial analysis of the spatial information about objects and information about their dynamic actions in the temporal dimension.Behaviour recognition may even rely more on the modeling of temporal information containing short-range and long-range motions,in contrast to computer vision tasks involving images that focus on understanding spatial information.However,current solutions fail to jointly and comprehensively analyse short-range motions between adjacent frames and long-range temporal aggregations at large scales in videos.In this paper,we propose a novel behaviour recognition method based on the integration of multigranular(IMG)motion features,which can provide support for deploying video analysis in multimedia IoT crowdsensing systems.In particular,we achieve reliable motion information modeling by integrating a channel attention-based short-term motion feature enhancement module(CSEM)and a cascaded long-term motion feature integration module(CLIM).We evaluate our model on several action recognition benchmarks,such as HMDB51,Something-Something and UCF101.The experimental results demonstrate that our approach outperforms the previous state-of-the-art methods,which confirms its effective-ness and efficiency.
基金National Key R&D Program of China,Grant/Award Number:2022YFC3300704National Natural Science Foundation of China,Grant/Award Numbers:62171038,62088101,62006023。
文摘Due to hardware limitations,existing hyperspectral(HS)camera often suffer from low spatial/temporal resolution.Recently,it has been prevalent to super-resolve a low reso-lution(LR)HS image into a high resolution(HR)HS image with a HR RGB(or mul-tispectral)image guidance.Previous approaches for this guided super-resolution task often model the intrinsic characteristic of the desired HR HS image using hand-crafted priors.Recently,researchers pay more attention to deep learning methods with direct supervised or unsupervised learning,which exploit deep prior only from training dataset or testing data.In this article,an efficient convolutional neural network-based method is presented to progressively super-resolve HS image with RGB image guidance.Specif-ically,a progressive HS image super-resolution network is proposed,which progressively super-resolve the LR HS image with pixel shuffled HR RGB image guidance.Then,the super-resolution network is progressively trained with supervised pre-training and un-supervised adaption,where supervised pre-training learns the general prior on training data and unsupervised adaptation generalises the general prior to specific prior for variant testing scenes.The proposed method can effectively exploit prior from training dataset and testing HS and RGB images with spectral-spatial constraint.It has a good general-isation capability,especially for blind HS image super-resolution.Comprehensive experimental results show that the proposed deep progressive learning method out-performs the existing state-of-the-art methods for HS image super-resolution in non-blind and blind cases.
基金supported by National Key Basic Research Program of China(973 Program) under Grant No.2014CB340404National Natural Science Foundation of China under Grant Nos.61272111 and 61273216Youth Chenguang Project of Science and Technology of Wuhan City under Grant No. 2014070404010232
文摘Point-of-interest(POI) recommendation is a popular topic on location-based social networks(LBSNs).Geographical proximity,known as a unique feature of LBSNs,significantly affects user check-in behavior.However,most of prior studies characterize the geographical influence based on a universal or personalized distribution of geographic distance,leading to unsatisfactory recommendation results.In this paper,the personalized geographical influence in a two-dimensional geographical space is modeled using the data field method,and we propose a semi-supervised probabilistic model based on a factor graph model to integrate different factors such as the geographical influence.Moreover,a distributed learning algorithm is used to scale up our method to large-scale data sets.Experimental results based on the data sets from Foursquare and Gowalla show that our method outperforms other competing POI recommendation techniques.
基金This study was supported by the National Defense Science and Technology Innovation Zone of China(Grant No.00205501).
文摘The reconstruction control of modular self-reconfigurable spacecraft (MSRS) is addressed using an adaptive sliding mode control (ASMC) scheme based on time-delay estimation (TDE) technology. In contrast to the ground, the base of the MSRS is floating when assembled in orbit, resulting in a strong dynamic coupling effect. A TED-based ASMC technique with exponential reaching law is designed to achieve high-precision coordinated control between the spacecraft base and the robotic arm. TDE technology is used by the controller to compensate for coupling terms and uncertainties, while ASMC can augment and improve TDE’s robustness. To suppress TDE errors and eliminate chattering, a new adaptive law is created to modify gain parameters online, ensuring quick dynamic response and high tracking accuracy. The Lyapunov approach shows that the tracking errors are uniformly ultimately bounded (UUB). Finally, the on-orbit assembly process of MSRS is simulated to validate the efficacy of the proposed control scheme. The simulation results show that the proposed control method can accurately complete the target module’s on-orbit assembly, with minimal perturbations to the spacecraft’s attitude. Meanwhile, it has a high level of robustness and can effectively eliminate chattering.
文摘The dynamic characteristics related to micro-motions, such as mechanical vibration or rotation, play an essential role in classifying and recognizing ballistic targets in the midcourse, and recent researches explore ways of extracting the micro-motion features from radar signals of ballistic targets. In this paper, we focus on how to investigate the micro-motion dynamic characteristics of the ballistic targets from the signals based on infrared (IR) detection, which is mainly achieved by analyzing the periodic fluctuation characteristics of the target IR irradiance intensity signatures. Simulation experiments demonstrate that the periodic characteristics of IR signatures can be used to distinguish different micro motion types and estimate related parameters. Consequently, this is possible to determine the micro-motion dynamics of ballistic targets based on IR detection.
基金supported by the Natural Science Foundation of China(No.61906213).
文摘Most of the current object detection algorithms use pretrained models that are trained on ImageNet and then fine-tuned in the network,which can achieve good performance in terms of general object detectors.However,in the field of remote sensing image object detection,as pretrained models are significantly different from remote sensing data,it is meaningful to explore a train-fromscratch technique for remote sensing images.This paper proposes an object detection framework trained from scratch,SRS-Net,and describes the design of a densely connected backbone network to provide integrated hidden layer supervision for the convolution module.Then,two necessary improvement principles are proposed:studying the role of normalization in the network structure,and improving data augmentation methods for remote sensing images.To evaluate the proposed framework,we performed many ablation experiments on the DIOR,DOTA,and AS datasets.The results show that whether using the improved backbone network,the normalization method or training data enhancement strategy,the performance of the object detection network trained from scratch increased.These principles compensate for the lack of pretrained models.Furthermore,we found that SRS-Net could achieve similar to or slightly better performance than baseline methods,and surpassed most advanced general detectors.
文摘Object detection in Remote Sensing(RS)has achieved tremendous advances in recent years,but it remains challenging for rotated object detection due to cluttered backgrounds,dense object arrangements and the wide range of size variations among objects.To tackle this problem,Dense Context Feature Pyramid Network(DCFPN)and a powerα-Gaussian loss are designed for rotated object detection in this paper.The proposed DCFPN can extract multi-scale information densely and accurately by leveraging a dense multi-path dilation layer to cover all sizes of objects in remote sensing scenarios.For more accurate detection while avoiding bottlenecks such as boundary discontinuity in rotated bounding box regression,a-Gaussian loss,a unified power generalization of existing Gaussian modeling losses is proposed.Furthermore,the properties ofα-Gaussian loss are analyzed comprehensively for a wider range of applications.Experimental results on four datasets(UCAS-AOD,HRSC2016,DIOR-R,and DOTA)show the effectiveness of the proposed method using different detectors,and are superior to the existing methods in both feature extraction and bounding box regression。
文摘In order to rapidly and accurately detect infrared small and dim targets in the infrared image of complex scene collected by virtual prototyping of space-based downward-looking multiband detection,an improved detection algorithm of infrared small and dim target is proposed in this paper.Firstly,the original infrared images are changed into a new infrared patch tensor mode through data reconstruction.Then,the infrared small and dim target detection problems are converted to low-rank tensor recovery problems based on tensor nuclear norm in accordance with patch tensor characteristics,and inverse variance weighted entropy is defined for self-adaptive adjustment of sparseness.Finally,the low-rank tensor recovery problem with noise is solved by alternating the direction method to obtain the sparse target image,and the final small target is worked out by a simple partitioning algorithm.The test results in various spacebased downward-looking complex scenes show that such method can restrain complex background well by virtue of rapid arithmetic speed with high detection probability and low false alarm rate.It is a kind of infrared small and dim target detection method with good performance.
文摘The electrical performance of radomes on high-speed aircraft can be influenced by the thermal and mechanical loads produced during high-speed flight,which can affect the detection dis-tance and accuracy of the guidance system.This paper presents a new method that uses the Finite Difference Time Domain(FDTD)method to calculate the electrical performance of radomes under Thermo-Mechanical-Electrical(TME)coupling.This method can accurately characterize the effects of material dielectric temperature drift and structural deformation on the electrical performance of the radome under flight conditions,enabling high-precision full-wave calculations of the broadband electrical performance of the radome.The method initiates by utilizing a Finite Element Grid Model(FE-GM)of the radome to sequentially acquire the radome's response temperature field and structural deformation field through thermal and mechanical simulations.Subsequently,spatial mapping techniques are developed to accurately incorporate the dielectric temperature drift and structural deformation of the radome into its Yee grid Electromagnetic(EM)simulation model.A verification case was designed to test the proposed method,and the results confirmed its high computational accuracy.Additionally,the effectiveness and necessity of the method were further demonstrated by analyzing the electrical performance of a fused silica ceramic radome used on a high-speed aircraft.
基金supported by the National Natural Science Foundation of China(71671059)
文摘The earth observation satellites(EOSs)scheduling problem for emergency tasks often presents many challenges.For example,the scheduling calculation should be completed in seconds,the scheduled task rate is supposed to be as high as possible,the disturbance measure of the scheme should be as low as possible,which may lead to the loss of important observation opportunities and data transmission delays.Existing scheduling algorithms are not designed for these requirements.Consequently,we propose a rolling horizon strategy(RHS)based on event triggering as well as a heuristic algorithm based on direct insertion,shifting,backtracking,deletion,and reinsertion(ISBDR).In the RHS,the driven scheduling mode based on the emergency task arrival and control station time window events are designed to transform the long-term,large-scale problem into a short-term,small-scale problem,which can improve the schedulability of the original scheduling scheme and emergency response sensitivity.In the ISBDR algorithm,the shifting rule with breadth search capability and backtracking rule with depth search capability are established to realize the rapid adjustment of the original plan and improve the overall benefit of the plan and early completion of emergency tasks.Simultaneously,two heuristic factors,namely the emergency task urgency degree and task conflict degree,are constructed to improve the emergency task scheduling guidance and algorithm efficiency.Finally,we conduct extensive experiments by means of simulations to compare the algorithms based on ISBDR and direct insertion,shifting,deletion,and reinsertion(ISDR).The results demonstrate that the proposed algorithm can improve the timeliness of emergency tasks and scheduling performance,and decrease the disturbance measure of the scheme,therefore,it is more suitable for emergency task scheduling.
基金supported by the Joint Fund of Equipment Pre-Research and Aerospace Science and Industry (6141B07090102)。
文摘For the detection of marine ship objects in radar images, large-scale networks based on deep learning are difficult to be deployed on existing radar-equipped devices. This paper proposes a lightweight convolutional neural network, LiraNet, which combines the idea of dense connections, residual connections and group convolution, including stem blocks and extractor modules.The designed stem block uses a series of small convolutions to extract the input image features, and the extractor network adopts the designed two-way dense connection module, which further reduces the network operation complexity. Mounting LiraNet on the object detection framework Darknet, this paper proposes Lira-you only look once(Lira-YOLO), a lightweight model for ship detection in radar images, which can easily be deployed on the mobile devices. Lira-YOLO's prediction module uses a two-layer YOLO prediction layer and adds a residual module for better feature delivery. At the same time, in order to fully verify the performance of the model, mini-RD, a lightweight distance Doppler domain radar images dataset, is constructed. Experiments show that the network complexity of Lira-YOLO is low, being only 2.980 Bflops, and the parameter quantity is smaller, which is only 4.3 MB. The mean average precision(mAP) indicators on the mini-RD and SAR ship detection dataset(SSDD) reach 83.21% and 85.46%, respectively,which is comparable to the tiny-YOLOv3. Lira-YOLO has achieved a good detection accuracy with less memory and computational cost.
文摘In the high speed target environment,there exists serious Doppler effect in the low pulse repetition frequency（LPRF） modulated frequency stepped frequency（MFSF） radar signal.The velocity range of the target is large and the velocity is high ambiguous,so the single method is difficult to satisfy the velocity measurement requirement.For this problem,a novel method is presented,it is a combination of cross-correlation inner frame velocity measurement and range-Doppler coupling velocity measurement.The cross-correlation inner frame method,overcoming the low Doppler tolerance of the cross-correlation between frames,can obtain the coarse velocity of the high speed target,and then the precision velocity can be obtained with the range-Doppler coupling method.The simulation results confirm the method is effective,and also it is well real-time and easy to the project application.
基金supported by the National High Technology Research and Development Program of China (863 Program) (2011AA1569)
文摘Satellite networking communications in navigation satellite system and spacebased deep space exploration have the features of a long delay and high bit error rate （BER）. Through analyzing the advantages and disadvantages of the Consulta tive Committee for the Space Data System （CCSDS） file delivery protocol （CFDP）, a new improved repeated sending file delivery protocol （RSFDP） based on the adaptive repeated sending is put forward to build an efficient and reliable file transmission. According to the estimation of the BER of the transmission link, RSFDP repeatedly sends the lost protocol data units （PDUs） at the stage of the retransmission to improve the success rate and reduce time of the retransmission. Theoretical analyses and results of the Opnet simulation indicate that the performance of RSFDP has significant improvement gains over CFDP in the link with a long delay and high BER. The realizing results based on the space borne filed programmable gate array （FPGA） platform show the applicability of the proposed algorithm.
基金supported by Natural Science Foundation of China (No. 62071466)
文摘Due to the attractive potential in avoiding the elaborate definition of anchor attributes,anchor-free-based deep learning approaches are promising for object detection in remote sensing imagery.Corner Net is one of the most representative methods in anchor-free-based deep learning approaches.However,it can be observed distinctly from the visual inspection that the Corner Net is limited in grouping keypoints,which significantly impacts the detection performance.To address the above problem,a novel and effective approach,called Group Net,is presented in this paper,which adaptively groups corner specific to the objects based on corner embedding vector and corner grouping network.Compared with the Corner Net,the proposed approach is more effective in learning the semantic relationship between corners and improving remarkably the detection performance.On NWPU dataset,experiments demonstrate that our Group Net not only outperforms the Corner Net with an AP of 12.8%,but also achieves comparable performance to considerable approaches with 83.4%AP.
基金supported in part by Beijing Natural Science Foundation(No.19L2002)in part by the National Natural Science Foundation of China(No.61631004)in part by BUPT Excellent Ph.D.students Foundation(No.CX2019312).
文摘Source localization plays an indispensable role in many applications.This paper addresses the directional source localization problem in a three-dimensional(3D)wireless sensor network using hybrid received-signal-strength(RSS)and angle-of-arrival(AOA)measurements.Both the position and transmission orientation of the source are to be estimated.In the considered positioning scenario,the angle and range measurements are respectively corresponding to the AOA model and RSS model that integrates the Gaussian-shaped radiation pattern.Given that the localization problem is non-convex and the unknown parameters therein are coupled together,this paper adopts the second-order cone relaxation and alternating optimization techniques in the proposed estimation algorithm.Moreover,to provide a performance benchmark for any localization method,the corresponding Cramer-Rao lower bounds(CRLB)of estimating the unknown position and transmission orientation of the source are derived.Numerical and simulation results demonstrate that the presented algorithm effectively resolves the problem,and its estimation performance is close to the CRLB for the localization with the hybrid measurements.
基金supported by the National Natural Science Foundation of China(61801513).
文摘Considering the sparsity of hyperspectral images(HSIs),dictionary learning frameworks have been widely used in the field of unsupervised spectral unmixing.However,it is worth mentioning here that existing dictionary learning method-based unmixing methods are found to be short of robustness in noisy contexts.To improve the performance,this study specifically puts forward a new unsupervised spectral unmixing solution.For the reason that the solution only functions in a condition that both endmembers and the abundances meet non-negative con-straints,a model is built to solve the unsupervised spectral un-mixing problem on the account of the dictionary learning me-thod.To raise the screening accuracy of final members,a new form of the target function is introduced into dictionary learning practice,which is conducive to the growing robustness of noisy HSI statistics.Then,by introducing the total variation(TV)terms into the proposed spectral unmixing based on robust nonnega-tive dictionary learning(RNDLSU),the context information under HSI space is to be cited as prior knowledge to compute the abundances when performing sparse unmixing operations.Ac-cording to the final results of the experiment,this method makes favorable performance under varying noise conditions,which is especially true under low signal to noise conditions.
基金supported by the Beijing Natural Science Foundation(4164097)the Chinese Postdoctoral Science Foundation(2016M591226)
文摘A new long term integration algorithm is proposed forhigh-dynamic targets, which can resolve the problems of spectrumspread, frequency walk and pseudorandom noise （PRN）code phase curvature caused by the motion of targets. This algorithmfirst applies a keystone transform based improved discretepolynomial-phase transform （KT-IDPT） to estimate the Dopplerchirp rate. Then, based on the estimated Doppler chirp rate,dechirping and envelope translation are performed on the partialcorrelation results to correct the spectrum spread and the codephase curvature. The simulation results demonstrate that the proposedmethod has low integration loss and computational burden.
基金Supported by the Ministerial Level Advanced Research Foundation of China(9140A05030213HT25012)
文摘Generally,Doppler fuze can only estimate actuation delay-time with a limited precision. As an improvement,imaging fuze can estimate actuation delay-time more precisely with the available two-dimensional image of the target. In this paper,imprecision of actuation delay-time estimation with Doppler fuze is first analyzed theoretically in brief. Secondly,feasibility analysis and theoretical model of imaging fuze are described,in which a criterion is established for the actuation delay-time based on the image,and then an image based gray-value weighted least square（ GWLS） algorithm is presented to calculate actuation delay-time of the imaging fuze. Finally,a simulation model of missiletarget near-field encounter is established. Simulation results indicate that actuation delay-time of the imaging fuze is estimated more precisely than by the Doppler fuze.
基金This work has been supported by the National Key R&D Program of China(No.2018YFB1802302)the National Natural Science Foundation of China(Nos.11774181,61727815,11274182,11904180,11804250 and 1190426)+2 种基金the Science and Technology Support Project of Tianjin(No.16YFZCSF00400)the Natural Science Foundation of Tianjin(No.19JCYBJC16700)the Tianjin Development Program for Innovation and Entrepreneurship。
文摘Laser communication is essential part of maritime-terrestrial-air intelligent communication/sensor network. Among them, different modulation formats would play a unique role in specific applications. Based on Rytov theory, we discussed system performance of the maritime laser communication with repeated coding technology in several modulation schemes. The closed-form expression of average bit error rate(BER) from weak to moderate atmospheric turbulence described by log-normal distribution is given. Differential phase shift keying(DPSK) modulation, as a potential solution for future maritime laser communication, has attracted a lot of attention. We analyzed the effects of atmospheric turbulence parameters(visibility, refractive index structure coefficient, non-Kolmogorov spectral power-law exponent, turbulence inner scale) and DPSK system parameters(receiver aperture diameter, repeat time) on average BER in detail. Compared with the aperture-averaging effects, the system BER can be well suppressed through increasing repeat time. This work is anticipated to provide a theoretical reference for maritime laser communication systems.
