An object tracking method based on guided filter for night fusion image

Online object tracking is a challenging problem as it entails learning an effective model to account for appearance change caused by intrinsic and extrinsic factors. In this paper, we propose a novel online object tracking with guided image filter for accurate and robust night fusion image tracking. Firstly, frame difference is applied to produce the coarse target, which helps to generate observation models. Under the restriction of these models and local source image, guided filter generates sufficient and accurate foreground target. Then accurate boundaries of the target can be extracted from detection results. Finally timely updating for observation models help to avoid tracking shift. Both qualitative and quantitative evaluations on challenging image sequences demonstrate that the proposed tracking algorithm performs favorably against several state-of-art methods.     

A categorization method of infrared polarization and intensity image fusion algorithm based on the transfer ability of difference features

The fusion of infrared polarization and intensity image can significantly improve the detection performance of target, and the fused image is more suitable for human visual perception and further image-processing tasks. In this paper, a new categorization method of infrared polarization and intensity image fusion algorithm based on the transfer ability of difference feature is proposed. Firstly, the difference feature between two kinds of image and the characteristics of different fusion algorithms are analyzed and summarized. Second, an evaluation vector of fusion algorithm for difference feature transform ability is constructed. Thirdly, the transfer ability of fusion algorithm for difference feature is estimated by the evaluation vector, and the degree of transfer ability of fusion algorithm for difference feature is analyzed. Finally the fusion algorithms are classified by the degree of transfer ability of fusion algorithm for difference feature. The results shows that the proposed fusion algorithm categorization method helps select fusion algorithms in actual scene.     

Guided filter and convolutional network based tracking for infrared dim moving target

The dim moving target usually submerges in strong noise, and its motion observability is debased by numerous false alarms for low signal-to-noise ratio. A tracking algorithm that integrates the Guided Image Filter (GIF) and the Convolutional neural network (CNN) into the particle filter framework is presented to cope with the uncertainty of dim targets. First, the initial target template is treated as a guidance to filter incoming templates depending on similarities between the guidance and candidate templates. The GIF algorithm utilizes the structure in the guidance and performs as an edge-preserving smoothing operator. Therefore, the guidance helps to preserve the detail of valuable templates and makes inaccurate ones blurry, alleviating the tracking deviation effectively. Besides, the two-layer CNN method is adopted to obtain a powerful appearance representation. Subsequently, a Bayesian classifier is trained with these discriminative yet strong features. Moreover, an adaptive learning factor is introduced to prevent the update of classifier’s parameters when a target undergoes sever background. At last, classifier responses of particles are utilized to generate particle importance weights and a re-sample procedure preserves samples according to the weight. In the predication stage, a 2-order transition model considers the target velocity to estimate current position. Experimental results demonstrate that the presented algorithm outperforms several relative algorithms in the accuracy.     

An improved POCS super-resolution infrared image reconstruction algorithm based on visual mechanism

The traditional projection onto convex sets (POCS) super-resolution (SR) reconstruction algorithm can only get reconstructed images with poor contrast, low signal-to-noise ratio and blurring edges. In order to solve the above disadvantages, an improved POCS SR infrared image reconstruction algorithm based on visual mechanism is proposed, which introduces data consistency constraint with variable correction thresholds to highlight the target edges and filter out background noises; further, the algorithm introduces contrast constraint considering the resolving ability of human eyes into the traditional algorithm, enhancing the contrast of the image reconstructed adaptively. The experimental results show that the improved POCS algorithm can acquire high quality infrared images whose contrast, average gradient and peak signal to noise ratio are improved many times compared with traditional algorithm.     

基于绝对差度量的灰阶图像光学匹配运算

讨论了基于绝对差度量的灰阶图像识别问题;利用光学相关原理和循环编码方法实现了基于绝对差度量的灰阶图像快速匹配处理;基于差异度矩阵和λ截集的概念来描述匹配结果,提高了灰阶目标识别抵抗各种畸变干扰的稳定性。此方法可以适用于任何的相干光和非相干光相关系统,给出了实验示例。     

基于P范数的核最小对数绝对差自适应滤波算法

为了进一步提高在a稳定分布噪声背景下非线性自适应滤波算法的收敛速度,本文提出了一种新的基于p范数的核最小对数绝对差自适应滤波算法(kernel least logarithm absolute difference algorithm based on p-norm, P-KLLAD).该算法结合核最小对数绝对差算法和p范数,一方面利用最小对数绝对差准则保证了算法在a稳定分布噪声环境下良好的鲁棒性,另一方面在误差的绝对值上添加p范数,通过p范数和一个正常数a来控制算法的陡峭程度,从而提高该算法的收敛速度.在非线性系统辨识和Mackey-Glass混沌时间序列预测的仿真结果表明,本文算法在保证鲁棒性能的同时提高了收敛速度,并且在收敛速度和鲁棒性方面优于核最小均方误差算法、核分式低次幂算法、核最小对数绝对差算法和核最小平均p范数算法.     

An infrared image super-resolution reconstruction method based on compressive sensing

Limited by the properties of infrared detector and camera lens, infrared images are often detail missing and indistinct in vision. The spatial resolution needs to be improved to satisfy the requirements of practical application. Based on compressive sensing (CS) theory, this thesis presents a single image super-resolution reconstruction (SRR) method. With synthetically adopting image degradation model, difference operation-based sparse transformation method and orthogonal matching pursuit (OMP) algorithm, the image SRR problem is transformed into a sparse signal reconstruction issue in CS theory. In our work, the sparse transformation matrix is obtained through difference operation to image, and, the measurement matrix is achieved analytically from the imaging principle of infrared camera. Therefore, the time consumption can be decreased compared with the redundant dictionary obtained by sample training such as K-SVD. The experimental results show that our method can achieve favorable performance and good stability with low algorithm complexity.     

基于多方向多尺度变换的图像压缩算法

针对小波变换方向选择性差的局限,提出了一种多方向多尺度的的图像变换。圆对称滤波器组首先将图像分解为高频子带和低频子带,然后利用方向滤波器组将高频子带分解为多个方向子带,而对低频子带进行小波变换。多方向多尺度变换能以更稀疏的方式表示图像的边缘和纹理等几何特征,有利于图像压缩。在该变换基础上,结合迭代量化、嵌入式块截断编码(EBCOT)和集合分裂嵌入式块编码(SPECK)构建一种压缩算法。实验结果表明,对于纹理和边缘丰富的图像,压缩算法的性能相对于JPEG2000有明显地提高。     

An absolute phase to world coordinates method for the defocusing structured light shape measurement system

This paper proposes a flexible calibration method for the defocusing structured light three-dimensional (3D) imaging system. Neither a high accuracy translation stage nor the parameters of the projector is required. Therefore the method is more flexible in many practical applications. World coordinates for calibration are obtained based on the method of coordinates from known plane. The monotonic nonlinear relationships between the absolute phase and each world coordinate have been investigated. Different from other methods, in our method, the 3D shape is directly recovered by the absolute phase to world coordinates method. Experimental results show that, when polynomial curve fitting methods are used, the root-mean-squared (RMS) error for the flat plate reconstruction is less than 0.12 mm, and the measurement error of 3D points is less than −0.34%. Furthermore, when comparing the measurement results of our method with the method of coordinates from known plane based on camera calibration, the relative error is less than −0.05%. Therefore, camera calibration is considered as the major error source.     

一种基于HSI颜色模型的目标提取方法

在工业、农业、军事以及医疗等领域中,目标识别是一个关键技术,近年来在图像处理方面已成为一个重要的研究课题。提出了一种用于除草机器人的基于HSI(色调、明度和饱和度)颜色模型的彩色图像目标提取方法。该方法首先将具有某种色彩信息的杂草与农作物从背景土壤中分离出来,然后再根据杂草与农作物形状的不同,采用腐蚀和膨胀的方法进一步将杂草与农作物分离。该方法除了用于除草机器人以外,还可用于水果采摘等。仿真结果验证了该方法的有效性。     

DGM-FD: A finite difference scheme based on the discontinuous Galerkin method applied to wave propagation

In this paper we formulate a numerical method that is high order with strong accuracy for numerical wave numbers, and is adaptive to non-uniform grids. Such a method is developed based on the discontinuous Galerkin method (DGM) applied to the hyperbolic equation, resulting in finite difference type schemes applicable to non-uniform grids. The schemes will be referred to as DGM-FD schemes. These schemes inherit naturally some features of the DGM, such as high-order approximations, applicability to non-uniform grids and super-accuracy for wave propagations. Stability of the schemes with boundary closures is investigated and validated. Proposed scheme is demonstrated by numerical examples including the linearized acoustic waves and solutions of non-linear Burger’s equation and the flat-plate boundary layer problem. For non-linear equations, proposed flux finite difference formula requires no explicit upwind and downwind split of the flux. This is in contrast to existing upwind finite difference schemes in the literature.     

A laser pistonphone based on self-mixing interferometry for the absolute calibration of measurement microphones

A laser pistonphone for the absolute calibration of microphones at low frequencies has been developed at UME. The motion of an electro-dynamically driven piston in a small close cavity produces a sound pressure. Accurate measurement of the piston displacement, by self-mixing interferometry, enables this sound pressure to be calculated, and consequently the pressure sensitivity of a microphone, exposed to this sound pressure, to be determined. Absolute calibrations of type LS1P and WS1P microphones have been carried out with an uncertainty of less than 0.15 dB. The performance of the laser pistonphone has been validated by comparing the measured microphone sensitivities with those obtained by the closed coupler reciprocity method.     

一种基于后验概率差值的拒识算法

本文提出了一种孤立词语音识别系统中基于后验概率差值的拒识算法。研究了作为拒识特征时，输入词的后验概率和后验概率差值之间的区别，并将多层感知人工神经网络用于拒识特征的学习。相比现存的几种拒识算法，本算法几乎不需要额外的计算和存储量。当识别率为98．2％时，拒识率达到了95．4％。     

基于粒子群优化算法的自适应图像分割方法

提出了一种基于粒子群优化算法的图像分割新方法。粒子群优化(PSO)算法是一类随机全局优化技术,它通过粒子间的相互作用发现复杂搜索空间中的最优区域缩短了寻找阈值的时间。将PSO用于基于改进的最佳加权熵阈值法的图像分割中,试验结果表明,该方法不仅能够避免陷入局部极值,而且其速度得到了明显的改善,是一种有效的图像分割新方法。     

基于畸变率的图像几何校正

大视场成像光学系统中的畸变会降低图像质量，必须预以校正。提出一种新的校正方法，即根据畸变率的定义推导出畸变校正公式。根据公式，在镜头畸变率已知的情况下可以很容易地校正畸变。对于畸变率未知的情况，给出了建立畸变模型的方法，通过畸变模型可近似计算畸变率。得出通过控制畸变模型中某一个形状的参数可以控制畸变量大小的结论。提出的方法已经在实际工程中采用。实践证明，这种模型可以满足大多数镜头的畸变校正要求。     

A mass-conserved multiphase lattice Boltzmann method based on high-order difference

The Z–S–C multiphase lattice Boltzmann model [Zheng, Shu, and Chew(ZSC), J. Comput. Phys. 218, 353(2006)]is favored due to its good stability, high efficiency, and large density ratio. However, in terms of mass conservation, this model is not satisfactory during the simulation computations. In this paper, a mass correction is introduced into the ZSC model to make up the mass leakage, while a high-order difference is used to calculate the gradient of the order parameter to improve the accuracy. To verify the improved model, several three-dimensional multiphase flow simulations are carried out,including a bubble in a stationary flow, the merging of two bubbles, and the bubble rising under buoyancy. The numerical simulations show that the results from the present model are in good agreement with those from previous experiments and simulations. The present model not only retains the good properties of the original ZSC model, but also achieves the mass conservation and higher accuracy.     

遗传算法应用于图像恢复的实验设计

结合遗传算法设计了图像恢复实验，分析了图像恢复的数学模型，给出具体的实验步骤和方法。     

Fast convergence and higher-order mode calculation of optical waveguides: perturbation method with finite difference algorithm

We present an efficient algorithm for determining mode eigenvalues as well as field distributions of optical waveguides with two-dimensional transverse refractive index profile. The algorithm is devised with the analytical perturbation correction method combined with the finite difference approximation of Helmoltz's equation. The technique is simple and does not involve solving any eigenvalue equation or matrix formalism. The algorithm reduces abruptly the computation time required for the field convergence to mode, and can calculate any higher-order modes without the need of any pre-conditioning the field w.r.t. waveguide geometry, or calculation of previous order modes and/orthogonalization. The analysis can yield precisely both scalar and polarized modes. By applying it to waveguide problems whose solutions are otherwise known, the efficacy of the method has been established.     

The absolute flatness measurements of two aluminum coated mirrors based on the skip flat test

Two methods based on the skip flat test are proposed to understand the absolute flatnesses of two aluminum coated mirrors, one with a circular aperture of 69 mm diameter, and the other with an elliptical aperture whose major and minor axes are 60 mm and 42 mm, respectively. For the mirror with the circular aperture, three measurements including one cavity measurement and two skip flat measurements are conducted. The measuring results are 0.079λ (PV) and 0.019λ (RMS). For the mirror with the elliptical aperture, two cavity measurements and one skip flat test are used. The measured results are 0.059λ (PV) and 0.014λ (RMS). The results agree well with those obtained with the mirrors measured at normal incidence, using high precision transmission flat as the reference. The agreements of the results indicate that the new methods could be used as two reliable ways to understand the absolute flatnesses of the aluminum coated mirrors.     

时域有限差分法的混凝土超声波损伤分析

混凝土质量的好坏直接影响到建筑物的安全性,选用方便实用的方法对其内部裂纹进行检测具有显著意义。该文首先借助时域有限差分法,建立了混凝土声场数学模型。之后对不同内部缺陷尺寸的混凝土模型进行了超声波数值模拟,分别从时域和频域对接收到的超声信号进行了分析。结果发现,当内部存在缺陷时,接收信号时域幅值明显变小,接收信号的幅值与缺陷尺寸大小呈指数关系,且接收信号传播时间与缺陷宽度呈线性关系。最后以初次到达信号的频域能量为特征指标,得出其与混凝土内部开裂缺陷尺寸呈指数关系。该方法为混凝土超声无损检测建立合适的数学模型提供了参考。