基于CNN的灰度图像边缘提取算法中模板参数的研究  

Research on Template Parameters of Edge Detection of Gray-scale Images Based on Cellular Neural Networks

作　　者：黄蕾[1] 刘文波[1] 

机构地区：[1]南京航空航天大学自动化学院,江苏南京210016

出　　处：《计算机与现代化》2006年第6期69-71,74,共4页Computer and Modernization

摘　　要：边缘是图像的最基本的特征之一,边缘提取是图像分析中非常重要的初始步骤。本文根据细胞神经网络的状态方程和输出方程,推导出网络中任意一点的最终稳定状态和收敛过程中这点的状态值间的关系。基于灰度图像边缘提取的准则,确定了满足灰度图像边缘提取的细胞神经网络中模板参数的取值范围。在这个范围内,可以有效地提取灰度图像的边缘。最后给出了仿真结果,并对结果进行了定性的比较和分析。Edge is one of the basic characters ofimases, and edge detection is a very important step in image analysis. Based on the state equation and output equation, this paper proposes the ultimate stable state of each cell is related to the state of each cell the process. And based on the rule of edge detection of gray-scale images, this paper also propoees the range of parameters in template of CNN for detecting edges of gray-scale images. Edses of All gray-scale images can be detected on this range template. This paper lists the experiment result and also takes qualitative comparison and analysis to the result.

关 键 词：细胞神经网络 灰度图像 边缘提取 模板参数 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]