Discriminating the boundaries of half-tone images using the multiaperture gradient-estimate method

被引：0

作者：

Medennikov, PA ^{[1
]}

机构：

[1] All Russia Sci Ctr, SI Vavilov State Opt Inst, Sci Res Inst Comprehens Testing Optoelect Devices, Sosnovyi Bor, Leningrad Oblas, Russia

来源：

JOURNAL OF OPTICAL TECHNOLOGY | 2002年 / 69卷 / 01期

关键词：

Algorithms - Gradient methods - Mathematical operators - Spurious signal noise;

D O I：

10.1364/JOT.69.000032

中图分类号：

O43 [光学];

学科分类号：

070207 ; 0803 ;

摘要：

This paper presents an algorithm for discriminating contours on half-tone images. An adaptive multiaperture system is proposed for calculating the boundary attribute for the image points, which makes it possible to fine-tune to the type of boundaries of the objects and to the noise level on the image. The contour is discriminated by tracing the local maxima of the boundary attribute. The results of the operation of the given algorithm are compared with the results of the operation of its nonadaptive analog, as well as with the results of the discrimination of boundaries by means of an algorithm based on Sobel's differential operator. (C) 2002 Optical Society of America.

引用

页码：32 / 35

页数：4