基于仿射不变特征的图像匹配

图像匹配是计算机视觉中的一个重要研究领域，无论在民用还是军用上都有着重要的应用价值。本文以研究室国防重点预研究项目自动目标识别为背景，采用图像匹配方法，实现飞行器定位导航。具体工作流程是：事先利用侦察手段获取飞行器途经下方的地物景象(基准图)并存于飞行器载计算机中，然后当携带相应传感器的飞行器飞过预定的位置范围时，拍摄当地的地物景象(实时图)，将实时图和基准图在飞行器载计算机中进行匹配比较，可确定当前飞行器的准确位置，完成定位导航功能。 由于对同一场景使用相同或不同的传感器(成像设备)，以及在不同条件下(天候、照度、摄像位置和角度等)成像的复杂性和多样性等困难的存在，传统的相关匹配方法对上述困难的克服在方法原理上存在先天不足，所以无法胜任。故本文采用的方法是基于局部不变量特征的图像匹配。局部不变量特征因为能更灵活地描述图像，有效地处理图像复杂和遮挡问题，所以基于局部不变量特征的图像匹配方法对于视点的大变化，图像背景变化，以及目标场景识别等都有较好的效果。 基于局部不变量特征的图像匹配方法的步骤通常分为三部分：(1)用图像区域检测算子提取图像相关区域，(2)构造合适的特征描述区域，(3)选择特征相似度度量准则实现图像区域特征的匹配。本文详细研究了最大稳定极值区域 (MSER)方法，在此基础上进行了改进，具体工作如下：(1)利用高斯核函数对图像平滑采样，建立图像的高斯尺度空间，(2)在图像的高斯尺度空间中，利用MSER检测算子检测出图像在不同尺度下的所有仿射相关区域，(3)由于区域不规则，再用仿射不变的椭圆拟合并归一化，这时所有的区域仅存在旋转的不同，(4)用SIFT特征描述图像区域，得到所有区域的128维特征向量集。(5)采用欧式距离度量特征间的相似度，以最近邻和次近邻的比值作为特征匹配准则进行匹配。 本论文的主要研究工作在于把图像的高斯尺度空间引入到MSER算法中，进而大大改善了MSER算法对于图像的尺度变换、仿射变换以及图像模糊的性能。由于建立了高斯尺度空间，增加了MSER检测算子检测的范围，所以使得改进算法的性能得到了改善。论文第四章给出四组实验，分别为尺度变换，仿射变换，图像模糊和大视点变换。最后通过对匹配结果正确数量和错误数量的统计，论证了改进方法的性能要好于MSER算法。通过对算法复杂度的分析，得出虽然在改进算法引入了图像的高斯尺度空间，但是算法复杂度却并未增加，与MSER算法相同，为O(nloglogn)。

Robust Guaranteed Cost Control for Yaw Control of Helicopter

针对具有时变不确定性且不确定性界为椭球的线性系统提出了一种新的具有自适应机制的鲁棒保性能控制器设计方法。首先,引入一个具有可由自适应律在线调整的可调参数的目标模型,通过该参数来保证由目标模型与被控模型所获得的误差系统渐近稳定。结合保证目标模型稳定性的设计,最终形成保证闭环系统稳定且控制器增益仿射依赖于可调参数的鲁棒保性能跟踪控制器。应用于安装在试验平台上的小型直升机航向控制中,仿真试验表明了该方法的有效性。

输电线巡检机器人控制与实验研究

近年来，机器人的应用越来越广泛和深入，输电线巡检机器人是当前特种作业机器人的研究热点之一，具有广泛的应用前景和实用价值。本文的研究内容是围绕国家“863”计划支持项目“500KV超高压输电线巡检机器人的研究”展开的。本研究工作针对巡检机器人的关键控制问题，主要由三部分组成：设计了巡检机器人的体系结构，并应用离散事件理论对机器人的任务、行为和动作建模；对巡检机器人双轮同步驱动控制进行了分析，并应用奇异摄动理论设计了控制器；研究了基于单目视觉的输电线立体定位方法及通过视觉伺服完成机器人自主抓线控制。 第一，介绍了巡检机器人的作业环境，重点探讨了机器人机械系统和控制系统的设计与实现。在机械子系统中，详细介绍了巡检机器人的机构实现与越障方法。在控制系统中详细阐述了基于分层递阶的机器人控制系统硬件组成。另外介绍了供电系统、无线传输系统、传感系统的设计与实现。分析了输电线路周围的电磁环境，及其对机器人的影响，并根据分析结果完成了对机器人的电磁防护设计。 第二，开展了输电线巡检机器人体系结构及人机交互系统研究，针对巡检机器人工作特点设计了基于规划和感知行为的混合式体系结构。针对巡检机器人工作环境设计了以机器人为中心的人机交互方式。参考前人建立的离散事件动力系统的层次结构和并行结构，提出了顺序结构并证明了其无阻塞性、可控性和监控器存在性，并结合以上三种结构建立了巡检机器人作业行为的离散动力学模型，分别获得了任务层、行为层和动作层的监控器。 第三，进行了巡检机器人双轮驱动控制研究。巡检机器人双轮行走机构为过驱动系统，对双轮行走系统进行了运动学和动力学建模，将一行走轮设为主动轮另一行走轮设为从动轮。针对两行走轮之间弹性关节导致的控制中的振荡问题，采用奇异摄动理论将系统分为快慢两个子系统；针对巡检机器人系统参数的时变性采用PD自适应算法设计了慢系统控制器；应用最优控制理论设计了快系统控制器。仿真结果验证了该方法的有效性。 第四，进行了输电线视觉定位和视觉伺服抓线问题的研究。输电线巡检机器人的自主越障控制是实现机器人实用化的关键问题。为实现巡检机器人自主越障，采用视觉伺服控制机械手臂自动抓线。为提取输电线图像特征点，针对输电线投影图像特征改进了边缘提取算法，应用聚类算法提取了输电线上的像素点。提出在机械手运动过程中采用EKF（扩展卡尔曼滤波）来实现对输电线的立体定位。在分析了当前基于图像的视觉伺服研究现状，建立了基于图像雅克比矩阵的输电线视觉伺服抓线模型。针对非标定状况下图像雅可比矩阵中的不确定参数，应用I&I(Immersion Invariant)自适应算法来实现无标定图像视觉伺服。针对机器人的动力学不确定性，设计了模糊自适应控制器，并证明了稳定性。仿真验证该方法的有效性，实验验证了基于视觉伺服的抓线控制的有效性。

压力和铁含量对铁方镁石(Mg1-xFex)O(x<=0.25)弹性常数影响的第一性原理研究

By first principle methods based on density functional theory (DFT),the equation of state(EOS) and elastic constants of both periclase and ferropericlase are calculated. The pressure and iron doping effects on the elastic constants of ferropericlase are investigated systematically. Firstly, we calculate the elastic constants of periclase and compare the obtained results with experimental data and other theoretical calculations, which shows a encouraging consistence and demonstrates the practicability of first-principle methods. Secondly, by adding iron into periclase crystal model, we build up ferropericlase with iron contents ranging from 0% to 25% mole percent. The corresponding elastic constants are calculated in a large pressure range(0~120GPa). Emphatically, the strong correlation of 3d electrons in transitional elements, such as iron, is difficult to treat in first-principle methods for a long time. The current solution is to make additional correction. During the initial stage of this study, the strong correlation of 3d electrons in iron is not considered, and we observed that addition of iron decreases the volume of ferropericlase, which is totally contradictory to the experimental data. By applying LDA+U approximation in order to solve the strongly correlated 3d electron of iron, we observed the expansion of volume by iron as expected. On the basis of the LDA+U approximation, the elastic constants of ferropericlase are calculated. After a detailed analysis of data obtained from theoretical calculations, we have reached the following conclusions:(1)pressure imposes positive effects on all elastic constants, and the degree of effects is C11>C12>C44. (2) Iron has no distinctive effects on C11 and C12, although some fluctuations are observed around 60GPa. However, iron has obvious softening effects on C44 The softening effects on C44 are intensified as pressure increases. Above the 100GPa, the effects increase greatly, even surpasses the pressure's positive effects in ferropericlase crystal models with iron mole percent of having 12.5%, 18.75% and 25% iron content. (3)As to the modulus deprived from elastic constants, iron has no effect on the adiabatic bulk module BS, only a little fluctuation around 60GPa. We find iron's softening effects on shear modulus G. (4)We find out that, compared with low iron content, elastic constants with iron content approaching 25mole% is consistently fluctuated，which may be caused by the limitations of the LDA+U approximation method itself. (5)We investigate the pressure and Fe doping effects on elastic anisotropy factor（A=(2C44+C12-C11)/C11） of ferropericlase and find out that iron contents will lower the critical isotropic pressure. At the same pressure, when the pressure is below the isotropic pressure, iron softens the anisotropy factor ; when pressure surpasses the isotropic pressure, iron increases the anisotropy factor.

Heat capacity enhancement and thermodynamic properties of nanostructured amorphous SiO2

The heat capacity of nanostructured amorphous SiO2 (na-SiO2) has been measured by adiabatic calorimetric method over the temperature range 9-354 K. TG and differential scanning calorimeter (DSC) were also employed to determine the thermal stability. Glass transition temperature (T-g) for the two same grain sizes with different specific surface of naSiO(2) samples and one coarse-grained amorphous SiO2 (ca-SiO2) sample were determined to be 1377, 1397 and 1320 K, respectively. The low temperature experimental results show that there are significant heat capacity (C-P) enhancements among na-SiO2 samples and ca-SiO2. Entropy, enthalpy, Gibbs free energy and Debye temperature (theta (D)) were obtained based on the low temperature heat capacity measurement of na-SiO2. The Cp enhancements of na-SiO2 were discussed in terms of configurational and vibrational entropy. (C) 2001 Elsevier Science B.V. All rights reserved.

Enhancement of molar heat capacity of nanostructured Al2O3

The nanosized alumina prepared by the hydrolysis method with an average particle size of 20 nm was characterized by X-ray diffraction. The heat capacity measurements of the prepared sample were carried out using an adiabatic calorimeter in the temperature range from 78 to 370 K. Enhancement of heat capacity was observed in the nanostructured materials as the heat capacity data were compared with those of the corresponding coarse-grained materials. The enhanced heat capacity was discussed on the basis of experiments. Differential scanning calorimetry and thermogravimetry were used to determine the thermal stability of the nanostructured alumina.

Low temperature specific heat and thermal stability of Fe-B ultrafine amorphous alloy particles

Fe-B ultrafine amorphous alloy particles (UFAAP) were prepared by chemical reduction of Fe3+ with NaBHO4 and confirmed to be ultrafine amorphous particles by transmission electron microscopy and X-ray diffraction. The specific heat of the sample was measured by a high precision adiabatic calorimeter, and a differential scanning calorimeter was used for thermal stability analysis. A topological structure of Fe-B atoms is proposed to explain two crystallization peaks and a melting peak observed at T=600, 868 and 1645 K, respectively.

Calorimetric study of Wood alloy

The heat capacities of Wood alloy have been measured with an automatic adiabatic calorimeter over the temperature range of 80 similar to 360 K. The thermodynamic data of solid-liquid phase transition have been obtained from the heat capacity measurements. The melting temperature, enthalpy and entropy of fusion of the substance are 345.662 K, 18.47 J.g(-1) and 0.05343 J.g(-1).K-1, respectively. The necessary thermal data are provided for the low temperature thermodynamic study of the alloy.

Heat capacity and thermochemical study of trifluoroacetamide (C2H2F3NO)

The low-temperature heat capacities of trifluoroacetamide were precisely determined with a small sample precision automated adiabatic calorimeter over the temperature range from 78 to 404 K. A solid-to-solid phase transition, a fusion and a phase transition from a liquid crystalline phase to fully liquid phase have been observed at the temperatures of 336.911+/-0.102, 347.622+/-0.094 and 388.896+/-0.160 K, respectively. The molar enthalpies of these phase transitions as well as the chemical purity of the substance were determined to be 5.576+/-0.004, 11.496+/-0.007, 1.340+/-0.005 kJ mol(-1) and 99.30 mol%, respectively, on the basis of the heat capacity measurements. The molar entropies of the three phase transitions were calculated to be 16.550+/-0.012, 33.071+/-0.029 and 3.447+/-0.027 J mol(-1) K-1, respectively. Further researches of the thermochemical properties for this compound have been carried out by means of TG and DSC techniques. (C) 2000 Elsevier Science B.V. All rights reserved.

Heat capacities of Al62.5Cu25Fe12.5 quasicrystals and B2 related crystals

The heat capacities of two Al62.5Cu25Fe12.5 samples containing icosahedral quasicrystals and B2 related crystals respectively were measured with a high-precision automatic adiabatic calorimeter over the temperature range of 75-385 K. The heat capacities of both samples increase with temperature. At the low temperature range, the heat capacity of the quasicrystalline sample is higher than that of the B2 approximate. However, the heat capacity of the B2 sample becomes higher above 254.987 K. (C) 1999 Elsevier Science B.V. All rights reserved.

Characterization and thermodynamic study of ultra-fine particle of Ni-B amorphous alloy

Ultra-fine particle of Ni-B amorphous alloy was prepared by chemical reduction of Ni2+ with NaBH4 and characterized with TEM and XRD. The heat capacity and thermal stability were measured with a high-precision automatic adiabatic calorimeter and DTA. The upper limit of applied temperature of the substance was found to be 684 K for use as catalyst. (C) 1999 Elsevier Science B.V. All rights reserved.

Low-temperature heat capacity and phase transition of n-hexatriacontane

The heat capacities of crystalline and liquid n-hexatriacontane were measured with an automatic adiabatic calorimeter over the temperature range of 80-370 K. Two solid-to-solid phase transitions at the temperatures of 345.397 and 346.836 K, and a fusion at the temperature of 348.959 K have been observed. The enthalpies and entropies of these phase transitions as well as the chemical purity of the substance were determined on the basis of the heat capacity measurements. Thermal decomposition temperatures of the compound were measured by thermogravimetric analysis. (C) 1999 Elsevier Science B.V. All rights reserved.

儿童汉字意识及精细辨认能力的发展研究

The important basis for the children to study chinese character is the Chinese character awareness, which means the understanding and extracting of the Chinese character combinatorial rule. The perceptual learning intrinsically is the direct perceive to the invariant rules. As the important mechanism to extract the abstract rule, the perceptual learning is the important mechanism of the development of the Chinese character awareness also. 4-6-year-old children's Chinese character awareness are tested by the degree of acceptance to the graphic words, inlaying words and pseudowords. The perceptual disintegration are tested by the perceptual learning experiment. Moreover, the development of the ability to finely-identify the Chinese character under the different level of attention condition is explored. The result suggests the quality to the development of the children's Chinese character awareness: The Chinese character awareness to 4-year-old children is in the rudiment stage, and it have different qualities as the change of the age. As the mechanism of the Chinese character awareness, the perceptual learning have an complex correlation to the awareness. The result indicates the following effects of the children perceptual disintegration: the level of the perceptual disintegration is transitive from three-dimension confusion to one-dimension falsity. It have correlation with the ability to finely-identify the Chinese character to a certain degree. The result also shows: the developmental specialties of the ability to finely-identify the Chinese character are different in the different age. The ability to finely-identify the Chinese character under the different level of attention condition have very similar processes.