On the Approximation of Denotational Mu-Semantics

A signature gives rise to a language L(Var) by extending with variables x Var and binding constructs x and x, corresponding to least and greatest fixed points respectively. The natural denotational models for such languages are bicomplete dcpos as monotone -algebras. We prove that several approximating denotational semantics have the usual compositional semantics as their limit. These results provide techniques for relating syntactic and semantic concepts such as in full abstraction or completeness proofs. In the presence of an involutive antitone map on a bicomplete dcpo D we may translate the language L(Var) into one with least fixed points only such that meanings are preserved. This allows an approximative semantics where least and greatest fixed points are simultaneously approximated by unwindings in the syntax, provided that the limit semantics is substitutive. We discuss the principal difficulties of simultaneous unwindings in the absence of such semantic negations.     

用递推法分析光学涂层的稳定性

应用递推方法得到了光学涂层稳定性的表达式,这些方程改进了涂层稳定条件的确定。应用这一方法,在掠角入射下对减反射涂层及全介质反射镜的稳定性进行了计算     

An integrable nonlinear diffusion hierarchy with a two-dimensional arbitrary function

By introducing the hereditary condition to a general first order differential strong symmetry operator, we obtain general 1+1 and 2+1 dimensional integrable nonlinear diffusion hierarchies with infinitely many symmetries and Lax pairs. For a special example the infinitely many nonlocal conservation laws and some explicit and implicit exact solutions are also given.     

一种基于区域搜索的全局运动估计与补偿算法

开展全局运动估计与补偿研究是进行动态目标检测中的基础和前提。在总结现有运动估计与补偿方法的基础上,提出一种基于图像分割区域的运动性和大小的全局运动估计与补偿算法。首先,通过建立区域搜索的全局运动模型,同时进行区域定性分割和区域大小排序;然后,根据误差最小化准则在指定的分割区域中进行线性递归搜索,利用门限准则寻找出最佳的运动估计参数;最后,根据双线性内插法获得运动位移量。实验结果比较可知,所提算法较三步搜索算法(TSS)和全局搜索算法(FSA)等传统算法具有更高的准确性(图像平移帧差)和实时性(算法运行时间),能够很好地实现运动背景的全局校正。