2020
2020
-
Record 217 of
Title:Deep Cross-Modal Image-Voice Retrieval in Remote Sensing
Author(s):Chen, Yaxiong(1,2); Lu, Xiaoqiang(1); Wang, Shuai(1)Source: IEEE Transactions on Geoscience and Remote Sensing Volume: 58 Issue: 10 DOI: 10.1109/TGRS.2020.2979273 Published: October 2020Abstract:With the rapid progress of satellite and aircraft technologies, cross-modal remote sensing image-voice retrieval has been studied in geography recently. However, there still exist some bottlenecks: how to consider the characteristics of remote sensing data adequately and how to reduce the memory and improve the retrieval efficiency in large-scale remote sensing data. In this article, we propose a novel deep cross-modal remote sensing image-voice retrieval approach, namely, deep image-voice retrieval (DIVR), to capture more information of remote sensing data to generate hash codes with low memory and fast retrieval properties. Especially, the DIVR approach proposes inception dilated convolution module to capture multiscale contextual information of remote sensing images and voices. Moreover, in order to enhance cross-modal similarity, the deep features' similarity term is designed to make paired similar deep features as close as possible and paired dissimilar deep features as mutually far as possible. In addition, the quantization error term is designed to drive hash-like codes to approximate hash codes, which can effectively reduce the quantization error for hash codes' learning. Extensive experimental results on three remote sensing image-voice data sets show that the proposed DIVR approach can outperform other cross-modal retrieval approaches. ? 1980-2012 IEEE.Accession Number: 20204209349066 -
Record 218 of
Title:Research on Initial Pointing of Inter-Satellite Laser Communication
Author(s):Jiaxin, Chen(1,2); Junfeng, Han(3)Source: Proceedings - 2020 12th International Conference on Intelligent Human-Machine Systems and Cybernetics, IHMSC 2020 Volume: 1 Issue: DOI: 10.1109/IHMSC49165.2020.00055 Published: August 2020Abstract:Laser communication has the advantages of low power consumption, small volume, large data transmission rate and so on.This technology has a broad application prospect. ATP(Acquisition,Tracking,Pointing) system is an important part of laser communication, in which the initial pointing plays a crucial role as the first step of acquisition. This paper establishes a mathematical model of initial pointing of inter-satellite laser communication, and by using MATLAB to simulate this mathematical model, the initial azimuth and pitch angle are obtained, and compared with the initial pointing angle obtained by STK(Satellite Tool Kit) under ideal conditions. The experimental results prove the correctness and feasibility of the mathematical model. ? 2020 IEEE.Accession Number: 20204409406833 -
Record 219 of
Title:Simulation Research of Non-line-of-sight Imaging System Based on Bidirectional Reflectance Distribution Function
Author(s):Xu, Wei-Hao(1,2); Su, Xiu-Qin(1); Wang, Shu-Chao(1,2); Zhu, Wen-Hua(1,2); Chen, Song-Mao(1,2); Wang, Ding-Jie(1,2); Wu, Jing-Yao(1,2)Source: Guangzi Xuebao/Acta Photonica Sinica Volume: 49 Issue: 12 DOI: 10.3788/gzxb20204912.1211002 Published: December 2020Abstract:The Non-Line-Of-Sight (NLOS) imaging process was studied to figure out the performance of existing NLOS algorithms under different reflection characteristics, with adopting physically based rendering bidirectional reflectance distribution function. Two state-of-the-art algorithms named f-k algorithm and Light-Cone Transform (LCT) algorithm are considered in the reconstruction using the proposed simulation system. The performance of the two algorithms are analyzed under various roughness, angles and niose. The simulation results show that: the change of reflection characteristics has a greater impact on the LCT algorithm; noise has a greater impact on the f-k algorithm. Based on the analysis of the experimental results, this article proposes an improvement to the f-k algorithm, merely using the phase information of the measured data for NLOS reconstruction. Improved algorithm is cpable to reconstruct target objects with different reflection characteristics, providing help for exploring further study. ? 2020, Science Press. All right reserved.Accession Number: 20210209739131 -
Record 220 of
Title:Design and Analysis of Hard X-Ray Microscope Employing Toroidal Mirrors Working at Grazing-Incidence
Author(s):Cui, Ying(1,2,3); Yan, Yadong(1); Wu, Bingjing(1); Li, Qi(1); He, Junhua(1)Source: International Journal of Pattern Recognition and Artificial Intelligence Volume: 34 Issue: 4 DOI: 10.1142/S0218001420550101 Published: April 1, 2020Abstract:A high resolution microscope is designed for plasma hard X-ray (10-20keV) imaging diagnosis. This system consists of two toroidal mirrors, which are nearly parallel, with an angle twice that of the grazing incidence angle and a plane mirror for spectral selection and correction of optical axis offset. The imaging characteristics of single toroidal mirror and double mirrors are analyzed in detail by the optical path function. The optical design, parameter optimization, image quality simulation and analysis of the microscope are carried out. The optimized hard X-ray microscope has a resolution better than 5μm at 1mm object field of view. The experimental data shows that the variation of the resolution is smaller in the direction of incident angle decrease than that in the increasing direction. ? 2020 World Scientific Publishing Company.Accession Number: 20193707419550 -
Record 221 of
Title:Generation of non-Kolmogorov atmospheric turbulence phase screen using intrinsic embedding fractional Brownian motion method
Author(s):Wang, Kaidi(1,2); Su, Xiuqin(1); Li, Zhe(1); Wu, Shaobo(1,2); Zhou, Wei(3); Wang, Rui(1,2); Chen, Songmao(1,2); Wang, Xuan(1,2,4)Source: Optik Volume: 207 Issue: DOI: 10.1016/j.ijleo.2020.164444 Published: April 2020Abstract:Generating phase screens to replace phase fluctuation caused by atmospheric turbulence is essential for simulation of light propagation through the atmosphere. Error between power spectral density of actual turbulence and traditional Kolmogorov model illustrates the importance of generating non-Kolmogorov phase screen. Meanwhile, methods used to generate phase screen at present show different kinds of disadvantages respectively. In this paper, we adopt a new method named "intrinsic embedding fractional Brownian motion (IE-FBM)". First, relationship between phase screen and FBM is analyzed. Next, principle of IE-FBM is clarified. We expand the correlation matrix and generate a stationary Gaussian surface through two fast Fourier transforms, which is the principle of intrinsic embedding. After that, we adjust the Gaussian surface into an FBM surface. Finally, simulation results demonstrate that IE-FBM combines advantages of traditional methods. Phase structure function becomes closer to theoretical value no matter how we set parameters of phase screen. Besides, both low and high frequency components of phase screen are sufficient and creases don't exist. In addition, time consumption reduces apparently. In conclusion, our method is comprehensively optimal choice to generate phase screen. ? 2020 Elsevier GmbHAccession Number: 20200908234852 -
Record 222 of
Title:Optical vortex with multi-fractional orders
Author(s):Hu, Juntao(1,2); Tai, Yuping(3); Zhu, Liuhao(1); Long, Zixu(1); Tang, Miaomiao(1); Li, Hehe(1); Li, Xinzhong(1,2); Cai, Yangjian(4,5)Source: Applied Physics Letters Volume: 116 Issue: 20 DOI: 10.1063/5.0004692 Published: May 18, 2020Abstract:Recently, optical vortices (OVs) have attracted substantial attention because they can provide an additional degree of freedom, i.e., orbital angular momentum (OAM). It is well known that the fractional OV (FOV) is interpreted as a weighted superposition of a series of integer OVs containing different OAM states. However, methods for controlling the sampling interval of the OAM state decomposition and determining the selected sampling OAM state are lacking. To address this issue, in this Letter, we propose a FOV by inserting multiple fractional phase jumps into whole phase jumps (2), termed as a multi-fractional OV (MFOV). The MFOV is a generalized FOV possessing three adjustable parameters, including the number of azimuthal phase periods (APPs), N; the number of whole phase jumps in an APP, K; and the fractional phase jump, α. The results show that the intensity and OAM of the MFOV are shaped into different polygons based on the APP number. Through OAM state decomposition and OAM entropy techniques, we find that the MFOV is constructed by sparse sampling of the OAM states, with the sampling interval equal to N. Moreover, the probability of each sampling state is determined by the parameter α, and the state order of the maximal probability is controlled by the parameter K, as K N. This work presents a clear physical interpretation of the FOV, which deepens our understanding of the FOV and facilitates potential applications, especially for multiplexing technology in optical communication based on OAM. ? 2020 Author(s).Accession Number: 20204209363188 -
Record 223 of
Title:Attribute-Cooperated Convolutional Neural Network for Remote Sensing Image Classification
Author(s):Zhang, Yuanlin(1); Zheng, Xiangtao(1); Yuan, Yuan(2); Lu, Xiaoqiang(1)Source: IEEE Transactions on Geoscience and Remote Sensing Volume: 58 Issue: 12 DOI: 10.1109/TGRS.2020.2987338 Published: December 2020Abstract:Remote sensing image (RSI) classification is one of the most important fields in RSI processing. It is well known that RSIs are very complicated due to its various kinds of contents. Therefore, it is very difficult to distinguish different scene categories with similar visual contents, like desert and bare land. To address hard negative categories, an attribute-cooperated convolutional neural network (ACCNN) is proposed to exploit attributes as additional guiding information. First, the classification branch extracts convolutional neural network feature, which is then utilized to recognize the RSI scene categories. Second, the attribute branch is proposed to make the network distinguish scene categories efficiently. The proposed attribute branch shares feature extraction layers with the classification branch and makes the classification branch aware of extra attribute information. Finally, the relationship branch constraints the relationship between the classification branch and the attribute branch. To exploit the attribute information, three attribute-classification data sets are generated (AC-AID, AC-UCM, and AC-Sydney). Experimental results show that the proposed method is competitive to state-of-the-art methods. The data sets are available at https://github.com/CrazyStoneonRoad/Attribute-Cooperated-Classification-Data sets. ? 1980-2012 IEEE.Accession Number: 20205009608642 -
Record 224 of
Title:Unsupervised variational auto-encoder hash algorithm based on multi-channel feature fusion
Author(s):Wang, Huanting(1,2); Qu, Bo(1); Lu, Xiaoqiang(1); Chen, Yaxiong(1,2)Source: Proceedings of SPIE - The International Society for Optical Engineering Volume: 11519 Issue: DOI: 10.1117/12.2573106 Published: 2020Abstract:Hashing technology is widely used to solve the problem of large-scale Remote Sensing (RS) image retrieval due to its high speed and low memory. Among the existing hashing algorithm, the unsupervised method is widely used in largescale RS image retrieval. However, the existing unsupervised RS image retrieval methods do not consider the multichannel properties of multi-spectral RS images and the discriminability in the local preservation mapping process adequately, which make it difficult to satisfy the retrieval performance of RS data. To solve these problems, we propose an unsupervised Variational Auto-Encoder Hashing algorithm based on multi-channel feature fusion (VAEH). MultiChannel Feature Fusion (MCFF) is used to extract the feature information of image, which fully considers the multichannel properties of the multi-spectral RS image. In order to enhance the discriminability in the local preservation mapping process, variational construction process and automatic encoder are added into the learning process of hashing function, and the KL distance of the Variational Auto-Encoder (VAE) is used to constrain the hashing code. Experiments on two large public RS image data sets (i.e. SAT-4 and SAT-6) have shown that our VAEH method outperforms the state of the art. ? 2020 SPIE.Accession Number: 20202908951759 -
Record 225 of
Title:Deep balanced discrete hashing for image retrieval
Author(s):Zheng, Xiangtao(1); Zhang, Yichao(1,2); Lu, Xiaoqiang(1)Source: Neurocomputing Volume: 403 Issue: DOI: 10.1016/j.neucom.2020.04.037 Published: 25 August 2020Abstract:Hashing has been widely used for large-scale multimedia retrieval because of its advantages in storage and retrieval efficiency. Traditional supervised hash methods represent an image as a feature vector and then perform a separate quantization step to generate a binary code. Due to the difficulty of discrete optimization of hash codes, continuous relaxation is generally used to replace discrete optimization. However, the process of continuous relaxation leads to inevitable quantization error. To avoid this drawback, a deep balanced discrete hashing method is proposed, which uses discrete gradient propagation with the straight-through estimator. The proposed method does not use the traditional continuous relaxation strategy, thereby reducing the quantization error caused by continuous relaxation. And the proposed method uses supervised information to directly guide the discrete coding and deep feature learning process. In the proposed method, the last layer of the Convolutional Neural Network (CNN) outputs the binary code directly. In the loss function, discrete values are calculated by combining the pairwise loss and a balance controlling term. The learned binary hash code maintains the similar relationship and label consistency at the same time. While maintaining the pairwise similarity, the proposed method keeps the balance of hash codes to improve retrieval performance. Extensive experiments show that the proposed method outperforms the state-of-the-art hashing methods on four image retrieval benchmark datasets. ? 2020 Elsevier B.V.Accession Number: 20202008665815 -
Record 226 of
Title:Research on Fuzzy Adaptive Control Algorithm with Extended Dimension for Disturbance Torque
Author(s):Changming, Lu(1); Xin, Gao(1); Meilin, Xie(2); Yu, Cao(3); Wei, Huang(2); Xuezheng, Lian(2); Kai, Liu(2); Wei, Hao(2)Source: Proceedings of 2020 IEEE 5th Information Technology and Mechatronics Engineering Conference, ITOEC 2020 Volume: Issue: DOI: 10.1109/ITOEC49072.2020.9141639 Published: June 2020Abstract:In order to solve the problem that friction, wire-wound, wind resistance and other disturbing moments seriously affect the stability tracking precision during the task of the photoelectric pod system, the fuzzy adaptive control algorithm with extended dimension is proposed in this paper. In this method, an accelerometer is first installed on the reflector of the pod. After obtaining the linear acceleration information and transforming it into angular acceleration, the fuzzy adaptive controller is designed according to the characteristics of wind resistance pulsation torque. The controller takes the mirror angular velocity, angular acceleration and target miss distance as input, and further adjusts the output of the controller according to the change of input and the fuzzy rule base of training. This algorithm was applied to the stable tracking experiment of a certain type of pod, and the results show that the tracking accuracy is improved from 59.7\mu\text{rad} to 32.4\ \mu\text{rad}. It is proved that the algorithm proposed in this paper can effectively suppress the disturbance torque and significantly improve the tracking accuracy and speed stability in the process of pod mission. This algorithm can be used in other servo control systems as a general method of disturbance torque suppression. ? 2020 IEEE.Accession Number: 20203809211553 -
Record 227 of
Title:Yb/Ce Codoped Aluminosilicate Fiber with High Laser Stability for Multi-kW Level Laser
Author(s):She, Shengfei(1); Liu, Bo(1); Chang, Chang(1); Xu, Yantao(1); Xiao, Xusheng(1); Cui, Xiaoxia(1); Li, Zhe(1); Zheng, Jinkun(1); Gao, Song(1); Zhang, Yan(1); Li, Yizhao(1); Zhou, Zhenyu(2); Mei, Lin(2); Hou, Chaoqi(1); Guo, Haitao(1)Source: Journal of Lightwave Technology Volume: 38 Issue: 24 DOI: 10.1109/JLT.2020.3019740 Published: December 15, 2020Abstract:Further power scaling and stable laser performance were demonstrated in the Yb/Ce codoped aluminosilicate fiber fabricated through low-temperature chelate gas phase deposition technique. The molar ratio of Ce/Yb was designed and optimized to be 0.58 for low background loss, effective photodarkening suppression, and no additional thermal load. The background loss of this active fiber was 4.7 dB/km and its photodarkening loss at equilibrium was as low as 3.9 dB/m at 633 nm. Benefiting from low-temperature deposition technique, the fiber showed uniform core composition devoid of clustering and central 'dip' of refractive index profile and 0.19 mol% Yb2O3 was homogeneously dissolved into the fiber core plus with 0.41 mol% Al2O3, 0.11 mol% Ce2O3, and 0.32 mol% SiF4. Based on a master oscillator power amplifier laser setup, 5.04 kW laser output at 1079.80 nm was achieved with a slope efficiency of 81.1%. Stabilized at 5kW-level laser for over 60 minutes, the output power presented almost no power degradation, directly confirming a noticeable photodarkening mitigation. ? 1983-2012 IEEE.Accession Number: 20205009615788 -
Record 228 of
Title:Exploiting Embedding Manifold of Autoencoders for Hyperspectral Anomaly Detection
Author(s):Lu, Xiaoqiang(1); Zhang, Wuxia(1,2); Huang, Ju(1,2)Source: IEEE Transactions on Geoscience and Remote Sensing Volume: 58 Issue: 3 DOI: 10.1109/TGRS.2019.2944419 Published: March 2020Abstract:Hyperspectral anomaly detection is an important task in the remote sensing domain. Recently, researchers have shown great interest in deep learning-based methods because they can learn hierarchical, abstract, and high-level representations. However, the latent features learned from the autoencoder (AE) are not always able to reflect the intrinsic structure of hyperspectral data because the locality property is not considered during the learning process. In order to address this problem, a novel manifold constrained AE network (MC-AEN)-based hyperspectral anomaly detection method is proposed in this article. First, the manifold learning method is employed to learn the embedding manifold. Then, the latent representations are learned by an AE network with the learned embedding manifold constraints to preserve the intrinsic structure of hyperspectral data. Finally, the reconstruction errors are calculated to detect anomalies. The global reconstruction error from MC-AEN and the local reconstruction error from the learned latent representations are combined to fully utilize the learned knowledge for better detection performance. We test our proposed algorithm on three different real data sets. Experimental results on these three data sets show the superiority of our proposed method. ? 1980-2012 IEEE.Accession Number: 20201108277661