2017
2017
-
Record 73 of
Title:Histogram-based human segmentation technique for infrared images
Author(s):Wu, Di(1,2); Zhou, Zuofeng(1); Yang, Hongtao(1); Cao, Jianzhong(1)Source: Advances in Intelligent Systems and Computing Volume: 555 Issue: DOI: 10.1007/978-981-10-3779-5_16 Published: 2017Abstract:Human detection in infrared video surveillance system is a challenging issue of computer vision. Effective human segmentation plays an important role in human detection. However, occlusion between different people makes it difficult to segment human groups. In this paper, we propose a new method for infrared human segmentation based on the histogram information. After selecting regions of interest with background subtraction, each connected human region is separated into single ones by analyzing histogram trend and calculating peak number. Experiment results show the accuracy of our method. ? 2017, Springer Nature Singapore Pte Ltd.Accession Number: 20173504103022 -
Record 74 of
Title:An improved non-uniformity correction algorithm and its hardware implementation on FPGA
Author(s):Rong, Shenghui(1); Zhou, Huixin(1); Wen, Zhigang(2,3); Qin, Hanlin(1); Qian, Kun(1); Cheng, Kuanhong(1)Source: Infrared Physics and Technology Volume: 85 Issue: DOI: 10.1016/j.infrared.2017.07.007 Published: September 2017Abstract:The Non-uniformity of Infrared Focal Plane Arrays (IRFPA) severely degrades the infrared image quality. An effective non-uniformity correction (NUC) algorithm is necessary for an IRFPA imaging and application system. However traditional scene-based NUC algorithm suffers the image blurring and artificial ghosting. In addition, few effective hardware platforms have been proposed to implement corresponding NUC algorithms. Thus, this paper proposed an improved neural-network based NUC algorithm by the guided image filter and the projection-based motion detection algorithm. First, the guided image filter is utilized to achieve the accurate desired image to decrease the artificial ghosting. Then a projection-based moving detection algorithm is utilized to determine whether the correction coefficients should be updated or not. In this way the problem of image blurring can be overcome. At last, an FPGA-based hardware design is introduced to realize the proposed NUC algorithm. A real and a simulated infrared image sequences are utilized to verify the performance of the proposed algorithm. Experimental results indicated that the proposed NUC algorithm can effectively eliminate the fix pattern noise with less image blurring and artificial ghosting. The proposed hardware design takes less logic elements in FPGA and spends less clock cycles to process one frame of image. ? 2017Accession Number: 20173404061253 -
Record 75 of
Title:Histograms of Gaussian normal distribution for feature matching in clutter scenes
Author(s):Zhou, Wei(1); Ma, Caiwen(1); Kuijper, Arjan(2)Source: arXiv Volume: Issue: DOI: Published: June 19, 2017Abstract:3D feature descriptors provide information between corresponding models and scenes. 3D objection recognition in cluttered scenes, however, remains a largely unsolved problem. Practical applications impose several challenges which are not fully addressed by existing methods. Especially in cluttered scenes there are many feature mismatches between scenes and models. We therefore propose Histograms of Gaussian Normal Distribution (HGND) for extracting salient features on a local reference frame (LRF) that enables us to solve this problem. We propose a LRF on each local surface patches using the scatter matrix’s eigenvectors. Then the HGND information of each salient point is calculated on the LRF, for which we use both the mesh and point data of the depth image. Experiments on 45 cluttered scenes of the Bologna Dataset and 50 cluttered scenes of the UWA Dataset are made to evaluate the robustness and descriptiveness of our HGND. Experiments carried out by us demonstrate that HGND obtains a more reliable matching rate than state-of-the-art approaches in cluttered situations. Copyright ? 2017, The Authors. All rights reserved.Accession Number: 20200034909 -
Record 76 of
Title:An effective method for human detection using far-infrared images
Author(s):Wu, Di(1); Wang, Jihong(1); Liu, Wei(2); Cao, Jianzhong(2); Zhou, Zuofeng(2)Source: 1st International Conference on Electronics Instrumentation and Information Systems, EIIS 2017 Volume: 2018-January Issue: DOI: 10.1109/EIIS.2017.8298602 Published: July 2, 2017Abstract:In this paper, a robust real-time approach to detect humans in far-infrared images is proposed. Adaptive thresholds and vertical edge operator are combined to extract human candidate regions. Then, disturbing components are removed using morphological operations, size filtering and component labeling. After analyzing each connected region through histogram evaluation, local thresholds are employed to separate overlapped human candidates into single ones. At last, nonhuman objects are eliminated by shape refinement. Experimental results demonstrate the approach is accurate to locate human regions and efficient to meet the real-time demand of a general surveillance system. ? 2017 IEEE.Accession Number: 20182605356031 -
Record 77 of
Title:Influence of Layup and Curing on the Surface Accuracy in the Manufacturing of Carbon Fiber Reinforced Polymer (CFRP) Composite Space Mirrors
Author(s):Yang, Zhiyong(1,2); Zhang, Jianbao(2); Xie, Yongjie(3); Zhang, Boming(1); Sun, Baogang(2); Guo, Hongjun(2)Source: Applied Composite Materials Volume: 24 Issue: 6 DOI: 10.1007/s10443-017-9595-7 Published: December 1, 2017Abstract:Carbon fiber reinforced polymer, CFRP, composite materials have been used to fabricate space mirror. Usually the composite space mirror can completely replicate the high-precision surface of mould by replication process, but the actual surface accuracy of replicated space mirror is always reduced, still needed further study. We emphatically studied the error caused by layup and curing on the surface accuracy of space mirror through comparative experiments and analyses, the layup and curing influence factors include curing temperature, cooling rate of curing, method of prepreg lay-up, and area weight of fiber. Focusing on the four factors, we analyzed the error influence rule and put forward corresponding control measures to improve the surface figure of space mirror. For comparative analysis, six CFRP composite mirrors were fabricated and surface profile of mirrors were measured. Four guiding control measures were described here. Curing process of composite space mirror is our next focus. ? 2017, Springer Science+Business Media Dordrecht.Accession Number: 20171003425216 -
Record 78 of
Title:Embedded measurement system of two-dimensional autocollimator based on FPGA
Author(s):Gao, Xiang(1); Hu, Xiaodong(2); Yang, Donglai(2); Zhang, Jian(3)Source: Proceedings of 2017 IEEE 3rd Information Technology and Mechatronics Engineering Conference, ITOEC 2017 Volume: 2017-January Issue: DOI: 10.1109/ITOEC.2017.8122304 Published: November 27, 2017Abstract:For the miniaturization of two-dimensional autocollimator, a method of using embedded measurement system instead of special host computer is presented. This system integrates CMOS image sensor's driving circuit, frame processing, adaptive exposure control, centroid subdivision and localization of cross, misalignment angle calculation, display driver and other functions within a FPGA chip, and the sampling image and measurement results are displayed through the TFTLCD mounted on the device body. The engineering prototype shows that the system has characters of high precision, high integration and high reliability. ? 2017 IEEE.Accession Number: 20181104893726 -
Record 79 of
Title:Research on video scene mapping of fixed viewing angle
Author(s):Wang, Yihao(1); Liu, Jiahang(1); Shi, Liu(1)Source: 2017 2nd International Conference on Image, Vision and Computing, ICIVC 2017 Volume: Issue: DOI: 10.1109/ICIVC.2017.7984599 Published: July 18, 2017Abstract:Mapping special images in video scene has practical important applications in the fields of advertising and television production, while there have been few reports on how to map in the background of video scene without impacting foreground targets which makes the result more realistic. We propose a method to embed images on certain location in video scene of fixed viewing angle. We first build background model from video frames, extract foreground using background subtraction method, then calibrate the camera using intrinsic information from video. On this basis we establish mapping matrices of image coordinate to world coordinate and image coordinate to video image coordinate according to location and orientation parameters. By using mapping matrices we embed the images on the background of video scene in right posture, and reproduce the foreground objects. Experiments in different scenes show that the proposed method is easily to use which makes mapping realistic and without impacting foreground objects, and has a good practicability. ? 2017 IEEE.Accession Number: 20173804169245 -
Record 80 of
Title:Properties analysis of composite materials for the manufacture of space mirror
Author(s):Yang, Zhiyong(1,2); Lei, Qin(2); Pan, Lingying(2); Tang, Zhanwen(2); Xie, Yongjie(3); Zhang, Boming(1); Sun, Jianbo(2); He, Xijun(2)Source: ICCM International Conferences on Composite Materials Volume: 2017-August Issue: DOI: Published: 2017Abstract:This work puts forward requirements of carbon fiber composite for space mirror, and compares properties of common intermediate modulus and high modulus carbon fibers and common resins of composites. Results show that carbon fiber composite for manufacturing space mirror should select high modulus carbon fiber and high toughness resin matrix. High toughness cyanate ester resin C705 and domestic high modulus carbon fiber were selected for manufacturing the prototype space mirror. ? 2017 International Committee on Composite Materials. All rights reserved.Accession Number: 20183705812596 -
Record 81 of
Title:Reweighted infrared patch-tensor model with both non-local and local priors for single-frame small target detection
Author(s):Dai, Yimian(1); Wu, Yiquan(1,2)Source: arXiv Volume: Issue: DOI: Published: March 27, 2017Abstract:Many state-of-the-art methods have been proposed for infrared small target detection. They work well on the images with homogeneous backgrounds and high-contrast targets. However, when facing highly heterogeneous backgrounds, they would not perform very well, mainly due to: 1) the existence of strong edges and other interfering components, 2) not utilizing the priors fully. Inspired by this, we propose a novel method to exploit both local and non-local priors simultaneously. Firstly, we employ a new infrared patch-tensor (IPT) model to represent the image and preserve its spatial correlations. Exploiting the target sparse prior and background non-local self-correlation prior, the target-background separation is modeled as a robust low-rank tensor recovery problem. Moreover, with the help of the structure tensor and reweighted idea, we design an entry-wise local-structure-adaptive and sparsity enhancing weight to replace the globally constant weighting parameter. The decomposition could be achieved via the element-wise reweighted higher-order robust principal component analysis with an additional convergence condition according to the practical situation of target detection. Extensive experiments demonstrate that our model outperforms the other state-of-the-arts, in particular for the images with very dim targets and heavy clutters. Copyright ? 2017, The Authors. All rights reserved.Accession Number: 20200011597 -
Record 82 of
Title:Emotional textile image classification based on cross-domain convolutional sparse autoencoders with feature selection
Author(s):Li, Zuhe(1,2); Fan, Yangyu(1); Liu, Weihua(3); Yu, Zeqi(2); Wang, Fengqin(2)Source: Journal of Electronic Imaging Volume: 26 Issue: 1 DOI: 10.1117/1.JEI.26.1.013022 Published: January 1, 2017Abstract:We aim to apply sparse autoencoder-based unsupervised feature learning to emotional semantic analysis for textile images. To tackle the problem of limited training data, we present a cross-domain feature learning scheme for emotional textile image classification using convolutional autoencoders. We further propose a correlation-analysis-based feature selection method for the weights learned by sparse autoencoders to reduce the number of features extracted from large size images. First, we randomly collect image patches on an unlabeled image dataset in the source domain and learn local features with a sparse autoencoder. We then conduct feature selection according to the correlation between different weight vectors corresponding to the autoencoder's hidden units. We finally adopt a convolutional neural network including a pooling layer to obtain global feature activations of textile images in the target domain and send these global feature vectors into logistic regression models for emotional image classification. The cross-domain unsupervised feature learning method achieves 65% to 78% average accuracy in the cross-validation experiments corresponding to eight emotional categories and performs better than conventional methods. Feature selection can reduce the computational cost of global feature extraction by about 50% while improving classification performance. ? 2017 SPIE and IS&T.Accession Number: 20170903403356 -
Record 83 of
Title:Reweighted Infrared Patch-Tensor Model with Both Nonlocal and Local Priors for Single-Frame Small Target Detection
Author(s):Dai, Yimian(1); Wu, Yiquan(1,2)Source: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Volume: 10 Issue: 8 DOI: 10.1109/JSTARS.2017.2700023 Published: August 2017Abstract:Many state-of-the-art methods have been proposed for infrared small target detection. They work well on the images with homogeneous backgrounds and high-contrast targets. However, when facing highly heterogeneous backgrounds, they would not perform very well, mainly due to: 1) the existence of strong edges and other interfering components, 2) not utilizing the priors fully. Inspired by this, we propose a novel method to exploit both local and nonlocal priors simultaneously. First, we employ a new infrared patch-tensor (IPT) model to represent the image and preserve its spatial correlations. Exploiting the target sparse prior and background nonlocal self-correlation prior, the target-background separation is modeled as a robust low-rank tensor recovery problem. Moreover, with the help of the structure tensor and reweighted idea, we design an entrywise local-structure-adaptive and sparsity enhancing weight to replace the globally constant weighting parameter. The decomposition could be achieved via the elementwise reweighted higher order robust principal component analysis with an additional convergence condition according to the practical situation of target detection. Extensive experiments demonstrate that our model outperforms the other state-of-the-arts, in particular for the images with very dim targets and heavy clutters. ? 2008-2012 IEEE.Accession Number: 20172203708621 -
Record 84 of
Title:Hardware in-loop system for X-ray pulsar-based navigation and experiments
Author(s):Zhang, Dapeng(1); Zheng, Wei(1); Sheng, Lizhi(2); Wang, Yidi(1); Xu, Neng(2)Source: Lecture Notes in Electrical Engineering Volume: 438 Issue: DOI: 10.1007/978-981-10-4591-2_45 Published: 2017Abstract:X-ray pulsar-based navigation uses natural objects, the neutron star, in space as the navigation signal source. The advantages of the method are navigation information is complete, and the reliability and autonomy are high. It is a research hot spot at present both at home and abroad. As a result of the X-ray signal from the pulsars is very weak, it cannot penetrate the thickset atmosphere. In order to validate the pulsar navigation algorithms closer to the real conditions on ground, the special Hardware in-Loop System should be used to do the experiments. This paper adopted the system "Tianshu-II" which is developed by National University of Defense Technology and Xi’an Institute of Optics and Precision Mechanics research institute. A series of X-ray pulsar-based navigation experiments are carried out. Experimental results show that the algorithms are reliable. They are verified to be effective in the hardware-in-the-loop simulation. ? Springer Nature Singapore Pte Ltd. 2017.Accession Number: 20172003680377