2015

• Record 13 of
  
  Title:Relative Orientation Dependent on Dual Quaternions
  
  Author(s):Sheng, Qing H.(1); Shao, Sa(1); Xiao, Hui(2); Zhu, Feng(3); Wang, Qing(1); Zhang, Bin(1)
  Source: Photogrammetric Record  Volume: 30  Issue: 151  DOI: 10.1111/phor.12111  Published: September 1, 2015  
  Abstract:A new approach to relative orientation based on dual quaternions is proposed. Dual quaternions are used to express a unified description of the relative position and orientation of two images in a stereopair. The coplanarity condition equation and its linearised model based on dual quaternions are established. According to the principle of least squares adjustment with constraints, a dual quaternion can be identified and the relative-orientation parameters can then be obtained by a non-linear transformation of the components of the dual quaternion. Experimental results show that the proposed approach is feasible and more reliable and efficient than the conventional approach based on Euler angles. © 2015 The Remote Sensing and Photogrammetry Society and John Wiley & Sons Ltd.
  Accession Number: 20153801282983
• Record 14 of
  
  Title:Study on the mutual alignment technology based on four-quadrant detectors
  
  Author(s):Hu, Ya-Bin(1,2); Wang, Miao(1)
  Source: Guangdianzi Jiguang/Journal of Optoelectronics Laser  Volume: 26  Issue: 11  DOI: 10.16136/j.joel.2015.11.0520  Published: November 15, 2015  
  Abstract:Mutual alignment accuracy is a main factor which affects the joint calibration of panoramic stereo cameras and laser radars in the dynamic spatial sensing area. To improve the result of joint calibration, we analyze the traditional technology of mutual alignment with telescope cross wires. As it is tedious and difficult, a new method of mutual alignment based on four-quadrant detectors is provided. The laser transmitter, four-quadrant detector and coarse aiming device are respectively arranged on two instruments. And the laser beam irradiates the active area of four-quadrant detector through coarse aiming. Then the center of light spot offset relative to the center of the active area is obtained according to the output voltage of four quadrants. Thus the positions and the poses of two instruments can be adjusted properly to realize mutual alignment. The simulation analysis and experimental results indicate that the four-quadrant detector alignment accuracy can meet the requirement of mutual alignment. This method provides a new idea for joint calibration, which operates more easily than the mutual alignment technology based on the telescope cross wire. © 2015, Board of Optronics Lasers. All right reserved.
  Accession Number: 20160801975734
• Record 15 of
  
  Title:A new design of filter system in streak camera
  
  Author(s):Zhou, Pengyu(1,2); Bai, Yonglin(1)
  Source: Proceedings of SPIE - The International Society for Optical Engineering  Volume: 9677  Issue:   DOI: 10.1117/12.2199534  Published: 2015  
  Abstract:In order to reduce the frequency of researchers routing in and out of the testing site and ensure the fluency of the testing we design a new filter system applied to the streak cameras. This system promotes streak cameras' abilities on spatial discrimination and time resolution. This paper focuses on the instruction of the piezoelectric motor's principle based on field-effect tubes. Filter wheel is driven by piezoelectric motor. It can effectively avoid the influences of high field produced by streak tube. Finally we achieve auto regulation at different gears and promote the efficiency of operations and guarantee the safety of researchers. CD4046 introduces the driven clock of this system and we use an inverter to get two synchronous inverted signals. These signals are amplified by field-effect tubes to more than 300V. The amplified ones are integrated at the output terminals to generate sinusoidal signal. The test shows that in this filter system piezoelectric motor operates at its resonance frequency under a control signal of 62.5 KHz. Its working current is 1.9A and driving power is almost 10W. By adjusting the gears, the filter wheel costs less than 2 seconds to calibrate. We accomplish the test in respected results. © COPYRIGHT SPIE 2015.
  Accession Number: 20155201714903
• Record 16 of
  
  Title:The design of aerial camera focusing mechanism
  
  Author(s):Hu, Changchang(1,2); Yang, Hongtao(1); Niu, Haijun(1)
  Source: Proceedings of SPIE - The International Society for Optical Engineering  Volume: 9677  Issue:   DOI: 10.1117/12.2199676  Published: 2015  
  Abstract:In order to ensure the imaging resolution of aerial camera and compensating defocusing caused by the changing of atmospheric temperature, pressure, oblique photographing distance and other environmental factor [1,2], and to meeting the overall design requirements of the camera for the lower mass and smaller size, the linear focusing mechanism is designed. Through the target surface support, the target surface component is connected with focusing driving mechanism. Make use of precision ball screws, focusing mechanism transforms the input rotary motion of motor into linear motion of the focal plane assembly. Then combined with the form of linear guide restraint movement, the magnetic encoder is adopted to detect the response of displacement. And the closed loop control is adopted to realize accurate focusing. This paper illustrated the design scheme for a focusing mechanism and analyzed its error sources. It has the advantages of light friction and simple transmission chain and reducing the transmission error effectively. And this paper also analyses the target surface by finite element analysis and lightweight design. Proving that the precision of focusing mechanism can achieve higher than 3um, and the focusing range is ±2mm. © 2015 SPIE.
  Accession Number: 20155201714946
• Record 17 of
  
  Title:Detecting small moving targets based on probability hypothesis density smoother
  
  Author(s):Li, Feipeng(1,2); Song, Zongxi(1); Li, Bin(1,2)
  Source: Journal of Information and Computational Science  Volume: 12  Issue: 14  DOI: 10.12733/jics20106698  Published: September 20, 2015  
  Abstract:In order to detect and track multiple small targets in low SNR environment, Probability Hypothesis Density (PHD) filter is proposed to solve the problem. When using the PHD filter to track small moving targets in space, the influence of measurement noise may be catastrophic. Measurement noise affects the calculation of particle weights, which results in the estimation error of targets number. This article brings in the concept of smoothing and combines it with PHD filter. When update particle weights, forward recursion and backward smoothing are both used. And at some extent, the influence of measurement noise is weakened. Finally through experiments, compared with standard PHD filter, the algorithm proposed in this article is found to be more accurate in targets number and state estimation. Copyright © 2015 Binary Information Press.
  Accession Number: 20155101706438
• Record 18 of
  
  Title:Fast and accurate circle detection based on trigonometric functions
  
  Author(s):Wu, Mengjie(1,2); Song, Zongxi(1); Gao, Wei(1); Li, Bin(1,2)
  Source: Journal of Information and Computational Science  Volume: 12  Issue: 10  DOI: 10.12733/jics20106137  Published: June 20, 2015  
  Abstract:Fast and accurate detection of circle is a challenging computer vision problem [1]. This paper takes fully advantages of circle good symmetry to calculate the midpoints of adjacent edge points in the horizontal direction. Based on the sine function reflected by the y coordinate values of midpoints and distance value l from the midpoints to the origin O and the cosine function reflected by the x coordinate values of midpoints and l, the midpoints are projected to parameter space. The main contributions of our work are threefold. 1) The problem of circle detection is formulated as sine curve and cosine curve detection. 2) The abundant phase information carried by trigonometric functions avoids the large amount calculation of Hough Transform. 3) A probabilistic scheme to recognize phase range is applied to circle detection. 1548-7741/Copyright © 2015 Binary Information Press
  Accession Number: 20153001050865
• Record 19 of
  
  Title:Automatic alignment method for diagnostic instrument
  
  Author(s):Wang, Wei(1); He, Junhua(1); Zhang, Min(1); Wei, Mingzhi(1); Yan, Yadong(1); Yang, Zhenghua(2); Yang, Pin(2)
  Source: Hongwai yu Jiguang Gongcheng/Infrared and Laser Engineering  Volume: 44  Issue: 3  DOI:   Published: March 25, 2015  
  Abstract:In order to capture accurate data of laser-driven fusion test, diagnostic instruments (DI) which are mounted onto diagnostic instrument manipulator (DIM) should align the target with tight tolerance. Traditional alignment methods have problems of time consumption and error large. An accurate automatic alignment method was introduced, which based on visual servoing according to the virtue of DIM and DI. First, a 3D alignment vector was constructed to estimate the target deviation in the stereo vision system. Under narrow field of view, estimated value was close to true deviation. Second, a 3 DOF pose adjusting mode was founded to promote motion accuracy. Finally, a visual servo controller based on the alignment vector and pose adjustment mode was designed. The process of alignment needs offline calibration only once. Through the three steps, DI can automatic align the target with high accuracy. Test indicates that alignment accuracy (RMS) is 11 μm in x direction and 12 μm in y direction. Laser shot test with X-ray framing camera show that the automatic align method can satisfy the project requirements. ©, 2015, Chinese Society of Astronautics. All right reserved.
  Accession Number: 20151800813116
• Record 20 of
  
  Title:Research of vehicle detection algorithm in infrared thermal imaging videos
  
  Author(s):Li, Bin(1,2); Wen, Desheng(1); Song, Zongxi(1); Shen, Chao(1,2); Wu, Mengjie(1,2); Li, Feipeng(1,2)
  Source: Journal of Information and Computational Science  Volume: 12  Issue: 9  DOI: 10.12733/jics20105946  Published: June 10, 2015  
  Abstract:This paper proposes a new algorithm for counting vehicle information in the infrared thermal imaging video sequences. First, the characteristics of the adjacent frames in video sequences has been analyzed, together with the modeling for the background information and updating background model according to the inter-frame variation. Second, the information of the moving target in the current frame is preliminarily extracted adopting the background subtraction algorithm, it achieves image segmentation and the division of the best detection area according to the imaging features of vehicles in different lanes by using morphological filtering and setting up the detection window. Then, model classification is done using the connected domain tags and template matching for vehicles in detection region. Finally, vehicle information each frame detects is counted by lanes. The experimental results present that the algorithm for dividing the best detection region segmentation image can reduce the miss rate and the false detection rate to the maximum extent. Meanwhile, the algorithm also has good detection accuracy for statistics of the traffic flow and classification of vehicles in video sequences. ©, 2015, Journal of Information and Computational Science. All right reserved.
  Accession Number: 20152801011031
• Record 21 of
  
  Title:A new effective image fusion algorithm based on NSCT and PCNN
  
  Author(s):Shen, Chao(1,2); Gao, Wei(1); Song, Zongxi(1); Yao, Tong(1,2); Li, Bin(1,2); Li, Feipeng(1,2); Wu, Mengjie(1,2)
  Source: Journal of Information and Computational Science  Volume: 12  Issue: 10  DOI: 10.12733/jics20106168  Published: June 20, 2015  
  Abstract:NonSubsampled Contourlet Transform (NSCT) has the characteristics of multi-scale, multi-directional, multi-resolution and shift invariance. Because of down sampling, traditional contourlet transform will cause Gibbs phenomenon, NSCT can overcome the disadvantage, obtaining better fusion image. Due to Pulse Coupled Neural Networks (PCNN) excellent biological characteristics, it has already been widely applied to image processing. In this paper, we will combining NSCT and PCNN, making full use of their advantages and applying to image fusion. The original image is by NSCT transform, getting the decomposition coefficients and then calculating its spatial frequency, input them to PCNN. According to the firing times select fusion coefficients. Experiment result has shown that, compared with the typical wavelet transform and contourlet transform method, the method proposed in this paper whether in subjective or objective is better than other methods. 1548-7741/Copyright © 2015 Binary Information Press
  Accession Number: 20153001050893
• Record 22 of
  
  Title:A new method of efficient measurement for sub-pixel image motion of space camera
  
  Author(s):Lu, Hao(1,2); Song, Zongxi(1); Gao, Wei(1); Yi, Hongwei(1); Wang, Qi(1,2); Xi, Jiangbo(1,2); Li, Bin(1,2)
  Source: Journal of Information and Computational Science  Volume: 12  Issue: 2  DOI: 10.12733/jics20105279  Published: January 20, 2015  
  Abstract:In order to measure the sub-pixel image motion of space camera which is caused by satellite attitude's instability or vibration, a method for image motion detection based on combination of corner detection and improved phase correlation is proposed in this paper. Firstly, it extracts feature points from the serial images using Harris detector. Then, the corners are used to construct a binary matrix. After that, the phase correlation for the constructed image is processed to get the relative output by taking the advantage of Fourier transform symmetry. Simultaneously, Blackman window function is taken to inhibit spectrum leakage caused by truncation for the output image. Finally, the sub-motion vector is calculated with centroid of mass calculation. On the basis of the theory proposed for measuring image motion, experiments for different sub-images with different image motions are carried out. Compared with other motion vector detection technology by soft algorithm, the simulations results show that this method not only inhibits the boundary effect but also avoids bringing truncation error, which improves the accuracy of estimation of image motion. Besides, this method effectively reduces the cost of computation and storage, improving the output SNR with good unimodality and real time property. The measure accuracy is good when the sub-image size is larger than 128128. The accuracy does not change with different scene of input images. The RMS error of image motion measurement can be controlled below 0.1 pixels, which meets the requirement of the space camera. Copyright © 2015 Binary Information Press.
  Accession Number: 20150500482277
• Record 23 of
  
  Title:Image quality evaluation method for compressive imaging system
  
  Author(s):Pan, Jinfeng(1,2,3); Qiu, Yuehong(1)
  Source: Journal of Computational Information Systems  Volume: 11  Issue: 15  DOI: 10.12733/jcis15039  Published: August 1, 2015  
  Abstract:Nowadays increased efforts have been made to devise digital still and high-speed video cameras designed based on the theory of compressive sensing. It is necessary to evaluate the image quality of compressive imaging systems during there design. An image quality evaluation method mainly based on the estimated modulation transfer function of the imaging system was proposed in this paper. The method estimates the distortions of the images captured by compressive imaging systems. The results of the experiments show that the proposed method is equivalent to the well-known image quality assessment method multi-scale structural similarity (MS-SSIM). Copyright © 2015 Binary Information Press.
  Accession Number: 20155201711788
• Record 24 of
  
  Title:Emif: Towards a scalable and effective indexing framework for large scale music retrieval
  
  Author(s):Shen, Jialie(1); Mei, Tao(2); Tao, Dacheng(3); Li, Xuelong(4); Rui, Yong(2)
  Source: ICMR 2015 - Proceedings of the 2015 ACM International Conference on Multimedia Retrieval  Volume:   Issue:   DOI: 10.1145/2671188.2749346  Published: June 22, 2015  
  Abstract:This article presents a novel indexing framework called EMIF (Effective Music Indexing Framework) to facilitate scalable and accurate content based music retrieval. EMIF system architecture is designed based on a "classification-and-indexing" principle and consists of two main functionality layers: 1) a novel semantic-sensitive classification to identify input music's category and 2) multiple indexing structures - one local indexing structure corresponds to one semantic category. EMIF's layered architecture not only enables superior search accuracy but also reduces query response time significantly. To evaluate the system, a set of comprehensive experimental studies have been carried out using large test collection and EMIF demonstrates promising performance over state-of-theart approaches. © Copyright 2015 ACM.
  Accession Number: 20161502216517