摘要
针对红外图像的特点,重点研究了采用双滤波的方法及侧抑制网络的“突出边框,增强反差”的功能,提出一种图像边缘检测方法,对噪声有一定的抑制作用,边缘检测果明显。还结合序列图像的相关性,运用免疫计算方法,分割出目标运动区域,形成目标模板,然后在运动序列的各帧动态更新模板。最后得到具有低对比度、低信噪比、边缘模糊的红外图像目标的序列边缘。
According to the properties of infrared image we have made a key research in double filtration and functions of highlighting frames and reinforcing contrast in lateral inhibition network, and put forward an image edge detection method. The method has some restraint over noise and evident effect in marginal detection. In light of sequence images' relevance we have utilized immune algorithm to segment object's moving areas and form object templates. Then we have obtained the updated templates in the state of moving sequence frames and finally low-contrast, low-noise and marginblurred sequence rims of infrared object images.
出处
《微电子学与计算机》
CSCD
北大核心
2007年第4期156-158,162,共4页
Microelectronics & Computer
关键词
侧抑制
免疫计算
边缘检测
lateral inhibition
immune algorithm
edge detection