复杂背景下扩展目标极形态分形分割算法

 用极形态算子优化分形模型，提出了一种复杂背景下的扩展目标分割新算法。首先分析极坐标系中可变尺度形态结构算子的旋转缩放特性，并以此为基础进行极形态运算，再利用最小二乘法提取分形尺度误差进行二值化，然后进行边界跟踪，依据先验知识抑制背景团块，最后保留并填充面积最大的目标团块。仿真实验证明能够有效分割复杂背景下的扩展目标，并较好地保留了目标形状特征。     

水分子凝聚体系的介孔结构分形学

透射电镜研究表明，4,40-双硬脂酰胺基二苯醚在水中聚集、自组装成缠绕细纤维状聚集体，进而使整个体系形成三维网络结构．水分子被包囊在这个网络结构中，形成一种新型的凝聚体系(水分子凝胶)．水分子凝胶是一种典型的纳米介孔物质，其复杂的微孔结构可以用分形维数D来表征，通过气体吸附方法(孔度法和比表面积法)计算，求得水分子凝胶体系的微孔结构的分形维数为2.1?2.2．对于纤维状三维网络结构的分形表征，通过粘度法和Cayley分形树模型得出分形维数为1.98．由此推测其分形网络形成的过程是一个初始成核-生长-枝化的循环过程．     

Generating synthetic time series from Bak-Sneppen co-evolution model mixtures

The Bak-Sneppen model of co-evolution is used to derive synthetic time series with a priori specified fractal dimension (or Hurst exponent) through a mixing of processes in various lattice dimensions. Both theoretical and numerical analyses concern the avalanches at the critical threshold and provide a model for time series reconstruction that can be tested as an alternative to the classical fractional Brownian motion (fBm) because of differences in properties. New results on critical threshold and avalanche structure are obtained up to Euclidean dimension d=6. The method involves a lattice-based structure and therefore is suitable for the application of parallel computing.     

基于复小波包分形理论的爬壁机器人故障检测

通过研究爬壁式机器人的控制和运动特征，提出一种基于复小波包分形理论的故障检测方法．利用复小波包的平移不变性，将爬壁式机器人传感器输出信号分解成独立的频谱，并进行阀值处理和重构，从而有效去除高频噪音并提取故障的特征频率；依据信号分形维数的多尺度不变性，在嵌入维数空间，采用维数最大距离法，确定复小波包域故障信号的关联雏数．仿真实验表明，爬壁式机器人在各种异常工作模式下的故障信号关联维数能比较真实地反映其故障状态空间，同时也验证了故障信号的关联维数低于正常信号的关联维数作为故障发生与否的定量判据的正确性．     

一种抗噪分形维计算方法及其在声纳图像识别中的应用

用于计算图像分形维的差分盒计数法(DBC)和鲁棒差分盒计数法(RDBC)都对脉冲噪声和斑点噪声较敏感,为此本文提出一种抗噪差分盒计数法(NRDBC),利用剪切局部标准差(TLSD)来计算图像分形维,由于TLSD可有效滤除脉冲噪声和斑点噪声、且对高斯噪声敏感性小,因此NRDBC能对有噪图像进行可靠的分形维估计。利用多分辨率的DBC、RDBC和NRDBC对混有高斯噪声、脉冲噪声和斑点噪声的7种Brodatz纹理以及3种海底的侧扫声纳图像进行了分类实验,结果表明,本文提出的NRDBC可获得更高的识别率和更好的抗噪性。     

The time dependent Ginzburg-Landau equation in fractal space-time

We use the hydrodynamic formulation of Scale Relativity Theory to analyze the TDGL equation. As a result, London equations come naturally from the system, when equating to zero the real velocity, the imaginary one turns real, the superconducting fluid act as a subquantum medium energy accumulator, the vector potential, the real and the imaginary velocity are all written in terms of the elliptic function. When solving the resulted system by means of WKBJ method, we get tunneling and quantization. In other words, scale transformation laws produce, on the motion equation of particles governed by the TDGL equation, under some peculiar assumptions, effects which are analogous to those of a “macroscopic quantum mechanics”.     

粉末注射成形坯孔隙度的分形模型

粉末注射成形坯是一种具有分形特性的典型的多孔介质,借助于多孔介质孔隙结构的分形理论,对粉末注射成形坯孔隙率的分形模型进行推导。首先分析了粉末注射成形坯孔隙结构的双重分形特性,介绍了粉末注射成形坯孔隙分布分形维数和孔隙迂曲分形维数,然后推导出粉末注射成形坯孔隙度的分形模型。     

The fractal dimension as estimator of the fractional content of metal matrix composite materials

Electrochemical techniques are applied to estimate the fractal dimension value of electroactive surface structures. However, the fractal dimension value is an abstract concept, which sometimes is hard to understand. Herein, this abstract concept is used to calculate the fractional content of the nickel/graphite–polypropylene hybrid composite material, putting into practice this concept in the study of composite materials.