一维统计量的椭圆检测法  

One-dimensional Statistics of the Detection Method

作　　者：夏耘[1] XIA Yun （Umversity of Shanghai For Science and Technology, Shanghai 200093, China）

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《电脑知识与技术》2009年第3期1710-1711,共2页Computer Knowledge and Technology

摘　　要：数字影像中目标的识别是十分关键的，该文以椭圆作为研究对象，在对于基于二维随机尖量的以HU不变矩为代表的不变矩算法上，提出了基于动差方法建立一维的随机统计量，并且证明该随机变量的三级动差和四级动差均为一常量。由于本算法只使用一维随机变量且该统计量又很容易获得，成为了其突现的优越性。Digital images of the .target recognition is crucial. This paper oval studied, based on two-dimensional random variables to the HU the same moment as represented by the same moment algorithm, based on the establishment of moments methods of one-dimenslonal random statistic, and that the three random variable Fixed-class and poor are four moments of a constant. As a result of this algorithm to use only one-dimensional random variables and the statistics are very easy to obtain, became the superiority of its emergence.

关 键 词：动差 识别 统计量 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论;自动化与计算机技术—计算机科学与技术]