摘要
提出了一种利用目前的CRS系统摄取六角网格数字图像的方法,并对获取的图像提出了基于金字塔处理的、从边缘检测等低层处理到模式识别等高层处理的算法.借助串行计算机模拟,利用六角几何和并行处理构架来识别六角网格数字图像中的物体.实验结果表明,该算法能够快速地识别一些简单的三维物体.最后提出了一个2层的六角网格视网膜的设计方案.文中论证了用于图像分析的六角像素网格的鲁棒性.
In this paper we introduce methods for acquiring hexagonal-array digital images through current CRS systems, and present pyramid-based processing algorithms from low level image processing such as edge detection to high level processing such as pattern recognition. Experimental tests of these algorithms by serial computer simulation on real digital images of simple 3-D objects, which makes maximum use of the hexagonal geometry and parallel computation architecture, have shown success at rapidly recognizing essential features of the objects boundaries. Finally we propose a design scheme for a two-layer hexagonalarray retina. This paper also investigates the robustness of hexagonal pixel grids for image analysis.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2007年第5期575-582,共8页
Journal of Computer-Aided Design & Computer Graphics
基金
中国博士后科学基金(2005037513).
关键词
六角网格
金字塔处理
并行计算
图像配准
边缘检测
特征提取
机器视觉
hexagonal grid
pyramid processing
parallel computation
image registration
edge detection
feature extraction
machine vision
