改进Bubble函数的多分辨率边缘提取方法  

Multi-Resolution Edge Extraction Way based on Improved Bubble Function

作　　者：余瑞星[1] 李言俊[1] 张科[1] 

机构地区：[1]西北工业大学航天学院,陕西西安710072

出　　处：《火力与指挥控制》2007年第7期8-10,共3页Fire Control & Command Control

基　　金：高等学校博士学科点专项科研基金(20020699014);国家航天基金资助项目(N4CH008)

摘　　要：在数学上,通常用Bubble函数来描述神经系统的侧抑制现象。Bubble函数在频域里相当于一带通滤波器,它可以反映不同的时域、频域特性,但它不具有调节带宽的功能。因此对Bubble函数进行改进,改进后的Bubble函数能够通过调节参数来获取不同时域、频域响应和滤波效果。最后,用改进的Bubble函数将小波和侧抑制有机地结合,构造出多分辨率侧抑制网络,该网络利用小波变换去除噪声、侧抑制网络突出边缘,从而可以准确、有效、快速地提取出图像边缘。In mathematics, Bubble function is always used to describe the lateral inhibition phenomenon of nerve system, and it can reflect the characters at time and frequency domains. Bubble function is a band pass filter in the frequency domains, while it can＇t have the ability to adjust bandwidth of filter. So an improved Bubble function was proposed, it can get different time-frequency domain response and filtering effect through adjusting parameter. Last, the improved Bubble function was used to combine Lateral inhibition with wavelet to get multi-resolution Lateral inhibition network. Using Lateral inhibition to enhance the contrast and wavelet threshold denoising to reduce the noise, we can distill the edge with lower noise exactly,effectively and quickly.

关 键 词：侧抑制 边缘提取 小波阈值去噪 离散小波变换 改进的Bubble函数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]