2021
2021
-
Record 181 of
Title:Multichannel left-subtract-right feature vector piston error detection method based on a convolutional neural network
Author(s):WANG, PENG-FEI(1,2); ZHAO, HUI(1); XIE, XIAO-PENG(1); ZHANG, YA-TING(1,2); LI, CHUANG(1); FAN, XUE-WU(1)Source: Optics Express Volume: 29 Issue: 14 DOI: 10.1364/OE.428690 Published: July 5, 2021Abstract:To realize the large-scale and high-precision co-phasing adjustment of syntheticaperture telescopes, we propose a multichannel left-subtract-right feature vector piston error detection method based on a convolutional neural network, which inherits the high precision and strong noise resistance of the DFA-LSR method while achieving a detection range of (-139λ, 139λ) (λ = 720 nm). In addition, a scheme to build large training datasets was proposed to solve the difficulty in collecting datasets using traditional neural network methods. Finally, simulations verified that this method can guarantee at least 94.96% accuracy with large samples, obtaining a root mean square error of 10.2 nm when the signal-to-noise ratio is 15. ? 2021 Optical Society of America.Accession Number: 20212610557854 -
Record 182 of
Title:Panoramic video motion small target detection algorithm in complex background
Author(s):Wang, Dian-Wei(1); Yang, Xu(1); Han, Peng-Fei(2); Liu, Ying(1); Xie, Yong-Jun(3); Song, Hai-Jun(3)Source: Kongzhi yu Juece/Control and Decision Volume: 36 Issue: 1 DOI: 10.13195/j.kzyjc.2019.0686 Published: January 2021Abstract:In order to solve the problem of low detection accuracy of moving small targets in the panoramic video in complex background, a small target detection algorithm based on complex background motion is proposed. Firstly, to reduce the interference of complex background information and improve the accuracy of target detection, the fast robust principal component analysis (Fast RPCA) algorithm is used to separate the foreground background information of the panoramic video image, and the foreground information is extracted as an effective image feature. Then, the candidate frame size of the region proposal network (RPN) in the faster region-convolutional neural networks (Faster R-CNN) is improved to adapt to the target size in the panoramic image, and then the foreground feature map is trained. Finally, the convolutional layer output detection model is shared by the RPN network and the Fast R-CNN network to achieve accurate detection of small targets in the panoramic video image. Experiments show that the proposed algorithm can effectively suppress the influence of complex background information on target detection accuracy, and has high detection accuracy for small moving targets in panoramic video images. Copyright ?2021 Control and Decision.Accession Number: 20210609884046 -
Record 183 of
Title:Broad-band phase retrieval method for transient radial shearing interference using chirp Z transform technique
Author(s):Xue, Fang(1,2); Duan, Ya-Xuan(1); Chen, Xiao-Yi(1,2); Li, Ming(1,2); Yuan, Suo-Chao(1,2); Da, Zheng-Shang(1)Source: Chinese Physics B Volume: 30 Issue: 8 DOI: 10.1088/1674-1056/abff2f Published: July 2021Abstract:The transient radial shearing interferometry technique based on fast Fourier transform (FFT) provides a means for the measurement of the wavefront phase of transient light field. However, which factors affect the spatial bandwidth of the wavefront phase measurement of this technology and how to achieve high-precision measurement of the broad-band transient wavefront phase are problems that need to be studied further. To this end, a theoretical model of phase-retrieved bandwidth of radial shearing interferometry is established in this paper. The influence of the spatial carrier frequency and the calculation window on phase-retrieved bandwidth is analyzed, and the optimal carrier frequency and calculation window are obtained. On this basis, a broad-band transient radial shearing interference phase-retrieval method based on chirp Z transform (CZT) is proposed, and the corresponding algorithm is given. Through theoretical simulation, a known phase is used to generate the interferogram and it is retrieved by the traditional method and the proposed method respectively. The residual wavefront RMS of the traditional method is 0.146λ, and it is 0.037λ for the proposed method, which manifests an improvement of accuracy by an order of magnitude. At the same time, different levels of signal-to-noise ratios (SNRs) from 50 dB to 10 dB of the interferogram are simulated, and the RMS of the residual wavefront is from 0.040λ to 0.066λ. In terms of experiments, an experimental verification device based on a phase-only spatial light modulator is built, and the known phase on the modulator is retrieved from the actual interferogram. The RMS of the residual wavefront retrieved through FFT is 0.112λ, and it decreases to 0.035λ through CZT. The experimental results verify the effectiveness of the method proposed in this paper. Furthermore, the method can be used in other types of spatial carrier frequency interference, such as lateral shearing interference, rotational shearing interference, flipping shearing interference, and four-wave shearing interference. ? 2021 Chinese Physical Society and IOP Publishing Ltd.Accession Number: 20213410799009 -
Record 184 of
Title:Optical design and analysis of compact visible and medium-wave infrared whisking broom imaging system
Author(s):Liu, Bo(1); Liu, Aimin(2); Li, Qiaoling(2); Xie, Laiyun(2)Source: Hongwai yu Jiguang Gongcheng/Infrared and Laser Engineering Volume: 50 Issue: 8 DOI: 10.3788/IRLA20200517 Published: August 25, 2021Abstract:To accomplish long-range visible and medium-wave infrared whisking broom imaging detection under strict space limitation, dual-band catadioptric shrink-beam system, double fast steering mirrors, and subsequet compact single-wave lenses was used to build a compact dual-band whisking broom imaging system through lens system design optimization. Among them, dual-band catadioptric shrink-beam system was composed of two-mirror Ritchey Chretien system, CaF2 dichroic prism and subsequet single-wave lenses. The image quality of the shrink-beam system was closed to diffraction limit in the 0.6-0.9 μm and 3.6-4.9 μm wave bands. Image motion of the dual-band shrink-beam system were controlled within halves of the respective pixels during broom imaging process. The effective focal length of the dual-band catadioptric system in the visible band was 1752 mm, there was no lens in between the RC, the three dimensional size of the optical system was 380 mm (axial)×Φ360, the telephoto ratio was 0.22, the line obscuration ratio was 0.34. Based on simulation and analysis, when the incident angles were larger than 30°, the point source transmittance (PST) of the dual-band system was less than 1×10?4 without additional front baffles. And this system was designed with mature optical cold working, installation and adjustment process and low cost. ? 2021, Editorial Board of Journal of Infrared and Laser Engineering. All right reserved.Accession Number: 20213610853430 -
Record 185 of
Title:Scattered Image Reconstruction at Near-infrared Based on Spatial Modulation Instability
Author(s):Liao, Yuan(1,2); Li, Lin(1,2); Wang, Zhaolu(1); Huang, Nan(1); Liu, Hongjun(1,3)Source: arXiv Volume: Issue: DOI: null Published: November 28, 2021Abstract:We present a method of near-infrared image reconstruction based on spatial modulation instability in a photorefractive strontium barium niobate crystal. The conditions that lead to the formation of modulation instability at near-infrared are discussed depending on the theory of modulation instability gain. Experimental results of scattered image reconstruction at the 1064 nm wavelength show the maximum cross-correlation coefficient and cross-correlation gain are 0.57 and 2.09 respectively. This method is expected to be an aid for near-infrared imaging technologies. Copyright ? 2021, The Authors. All rights reserved.Accession Number: 20210402357 -
Record 186 of
Title:Fine-Grained Visual Categorization by Localizing Object Parts with Single Image
Author(s):Zheng, Xiangtao(1); Qi, Lei(1); Ren, Yutao(2); Lu, Xiaoqiang(1)Source: IEEE Transactions on Multimedia Volume: 23 Issue: DOI: 10.1109/TMM.2020.2993960 Published: 2021Abstract:Fine-grained visual categorization (FGVC) refers to assigning fine-grained labels to images which belong to the same base category. Due to the high inter-class similarity, it is challenging to distinguish fine-grained images under different subcategories. Recently, researchers have proposed to firstly localize key object parts within images and then find discriminative clues on object parts. To localize object parts, existing methods train detectors for different kinds of object parts. However, due to the fact that the same kind of object part in different images often changes intensely in appearance, the existing methods face two shortages: 1) Training part detector for object parts with diverse appearance is laborious; 2) Discriminative parts with unusual appearance may be neglected by the trained part detectors. To localize the key object parts efficiently and accurately, a novel FGVC method is proposed in the paper. The main novelty is that the proposed method localizes the key object parts within each image only depending on a single image and hence avoid the influence of diversity between parts in different images. The proposed FGVC method consists of two key steps. Firstly, the proposed method localizes the key parts in each image independently. To this end, potential object parts in each image are identified and then these potential parts are merged to generate the final representative object parts. Secondly, two kinds of features are extracted for simultaneously describing the discriminative clues within each part and the relationship between object parts. In addition, a part based dropout learning technique is adopted to boost the classification performance further in the paper. The proposed method is evaluated in comparison experiments and the experiment results show that the proposed method can achieve comparable or better performance than state-of-the-art methods. ? 1999-2012 IEEE.Accession Number: 20202108688424 -
Record 187 of
Title:Bifurcation-tunable Diffractionless Light Propagation in One-dimensional Non-Hermitian Photonic Lattice
Author(s):Liu, Zhenjuan(1); Wang, Haohao(1); Dai, Yanan(1); Zhang, Zhiqing(1); Wang, Yishan(2); Qi, Xinyuan(1)Source: Guangzi Xuebao/Acta Photonica Sinica Volume: 50 Issue: 4 DOI: 10.3788/gzxb20215004.0423003 Published: April 25, 2021Abstract:The light bifurcation transmission without diffraction in one-dimensional periodic compound photonic lattice was studied theoretically and numerically. When the lattice equals to degenerated Su-Schrieffer-Heeger model, the incident light with the wave number k=±π will bifurcate into two symmetric branches without any diffraction and the angles between two beams can be controlled by the coupling J between two lattice sites. In addition, a modulation phase φ is introduced. When the non-Hermitian perturbations satisfy the parity-time symmetry, the diffractionless light bifurcation phenomenon with any incident light wave can be realized as long as the incident wave vector k and the modulation phase φ respect the expression k+φ=±π. Further studies have shown that the next-nearest coupling can control the transmission angle and power division of two branches. This research provides new ideas for the design of optical switches and future all-optical paths. ? 2021, Science Press. All right reserved.Accession Number: 20212010360078 -
Record 188 of
Title:Person Reidentification via Unsupervised Cross-View Metric Learning
Author(s):Feng, Yachuang(1); Yuan, Yuan(2); Lu, Xiaoqiang(1)Source: IEEE Transactions on Cybernetics Volume: 51 Issue: 4 DOI: 10.1109/TCYB.2019.2909480 Published: April 2021Abstract:Person reidentification (Re-ID) aims to match observations of individuals across multiple nonoverlapping camera views. Recently, metric learning-based methods have played important roles in addressing this task. However, metrics are mostly learned in supervised manners, of which the performance relies heavily on the quantity and quality of manual annotations. Meanwhile, metric learning-based algorithms generally project person features into a common subspace, in which the extracted features are shared by all views. However, it may result in information loss since these algorithms neglect the view-specific features. Besides, they assume person samples of different views are taken from the same distribution. Conversely, these samples are more likely to obey different distributions due to view condition changes. To this end, this paper proposes an unsupervised cross-view metric learning method based on the properties of data distributions. Specifically, person samples in each view are taken from a mixture of two distributions: one models common prosperities among camera views and the other focuses on view-specific properties. Based on this, we introduce a shared mapping to explore the shared features. Meanwhile, we construct view-specific mappings to extract and project view-related features into a common subspace. As a result, samples in the transformed subspace follow the same distribution and are equipped with comprehensive representations. In this paper, these mappings are learned in an unsupervised manner by clustering samples in the projected space. Experimental results on five cross-view datasets validate the effectiveness of the proposed method. ? 2013 IEEE.Accession Number: 20211310140065 -
Record 189 of
Title:Generation and Conversion Dynamics of Dual Bessel Beams with a Photonic Spin-Dependent Dielectric Metasurface
Author(s):Li, Tianyue(1); Li, Xingyi(2,3); Yan, Shaohui(2,3); Xu, Xiaohao(4); Wang, Shuming(1,5); Yao, Baoli(2,3); Wang, Zhenlin(1); Zhu, Shining(1,5)Source: Physical Review Applied Volume: 15 Issue: 1 DOI: 10.1103/PhysRevApplied.15.014059 Published: January 2021Abstract:A Bessel beam has the properties of propagation invariance and a self-healing effect, leading to a variety of interesting phenomena and applications. Recently, as a planar diffractive element with miniaturized size, metasurfaces are widely employed to manipulate light in the subwavelength region, including generating a Bessel beam. However, such a metasurface-generated Bessel beam allows output light with no tunable functions. Here, with the interplay of the geometric phase and the dynamic phase, we propose a method to generate and allow conversion from any orthogonal polarizations to independent Bessel beams with a single-layer dielectric metasurface. The simulation results indicate that the arbitrary conversion between different Bessel beams is related to the spin-dependent orbit motion caused by the tight-focusing effect, leading to the singularity of the spot. This physical mechanism is well studied and the theoretical model for revealing the dependence of different incident polarization on the conversion dynamics is presented. Our approach paves a way for efficient generation and multifunctional applications, ranging from high-numerical-aperture devices to compact nanophotonic platforms for spin-dependent structured beams. ? 2021 American Physical Society.Accession Number: 20210609894342 -
Record 190 of
Title:Improving physical parameters estimation in the single-beam multiple-intensity reconstruction
Author(s):Chen, Xiaoyi(1,2); Duan, Yaxuan(1); Da, Zhengshang(1)Source: Journal of Optics (United Kingdom) Volume: 23 Issue: 12 DOI: 10.1088/2040-8986/ac2ea9 Published: December 2021Abstract:The iterative phase retrieval based on phase diversity technologies can solve the stagnation problem of Gerchberg-Saxton algorithm which performs Fourier transform to iterate back and forth between the object and spectral planes with known constraints. However, the application of phase diversity technologies in iterative phase retrieval methods will bring in multiple physical parameters such as distances or wavelengths. The measured accuracy of these physical parameters will ultimately affect the accuracy of the iterative phase retrieval methods. In this paper, a physical parameters estimation method which has the advantages of high global convergence and local convergence is proposed to improve the accuracy of iterative phase retrieval methods. Meanwhile, this method is introduced in the single-beam multiple-intensity reconstruction (SBMIR), termed PE-SBMIR, and its performance is verified by simulations and experiments. By simulating multiple sets of distance parameters with errors, the retrieved accuracy using PE-SBMIR can be improved by 2-4 orders of magnitude compared with SBMIR. Experimental results show that whether it is an amplitude-type object or phase-type object, the accuracy using PE-SBMIR is significantly higher than using SBMIR. The physical parameters estimation method proposed in this paper may be adopted in other iterative phase retrieval methods using phase diversity technologies. ? 2021 IOP Publishing LtdAccession Number: 20214811231805 -
Record 191 of
Title:Non-iterative complex wave-field reconstruction based on Kramers-Kronig relations
Author(s):Shen, Cheng(1); Liang, Mingshu(1); Pan, An(2); Yang, Changhuei(1)Source: Photonics Research Volume: 9 Issue: 6 DOI: 10.1364/PRJ.419886 Published: June 1, 2021Abstract:A non-iterative and non-interferometric computational imaging method to reconstruct a complex wave field called synthetic aperture imaging based on Kramers-Kronig relations (KKSAI) is reported. By collecting images through a modified microscope system with pupil modulation capability, we show that the phase and amplitude profile of the sample at pupil limited resolution can be extracted from as few as two intensity images by using Kramers-Kronig (KK) relations. It is established that as long as each subaperture's edge crosses the pupil center, the collected raw images are mathematically analogous to off-axis holograms. This in turn allows us to adapt a recently reported KK-relations-based phase recovery framework in off-axis holography for use in KKSAI. KKSAI is non-iterative, free of parameter tuning, and applicable to a wider range of samples. Simulation and experiment results have proved that it has much lower computational burden and achieves the best reconstruction quality when compared with two existing phase imaging methods. ? 2021 Chinese Laser PressAccession Number: 20212310447580 -
Record 192 of
Title:High Precision Centroid Location Algorithm Based on Cubic Spline Fitting and Interpolation
Author(s):Liu, Jie(1,2); Zhang, Geng(1); Feng, Xiangpeng(1); Zhang, Zhinan(1); Li, Siyuan(1); Hu, Bingliang(1)Source: Guangxue Xuebao/Acta Optica Sinica Volume: 41 Issue: 12 DOI: 10.3788/AOS202141.1212004 Published: June 25, 2021Abstract:In this paper, we proposed a high-precision centroid location algorithm based on cubic spline fitting and interpolation combining the advantages of centroid algorithms and surface fitting algorithms. Besides, the principle behind its double error suppression was given. The simulation showed that the location error of the proposed algorithm at different SNRs was significantly smaller than that of traditional centroid algorithms. When the SNR was 20 dB, the root-mean-square error of the proposed algorithm was only 0.003 pixel. Furthermore, the star images from real instruments verified the effectiveness of this algorithm once again. In summary, this algorithm can effectively suppress the location error and has strong noise resistance, which can be widely applied without requiring specific star models. ? 2021, Chinese Lasers Press. All right reserved.Accession Number: 20213410810537