起伏背景的红外图像过渡区提取  

Transition region extraction of infrared image under clutter background

作　　者：刘锁兰[1] 杨静宇[1] 

机构地区：[1]南京理工大学计算机系,南京210094

出　　处：《计算机应用研究》2008年第2期501-503,共3页Application Research of Computers

基　　金：国防基础研究基金资助项目(J1500C002)

摘　　要：将基于过渡区提取的图像分割技术应用于起伏背景的红外图像的分割。为了实现复杂背景下过渡区的准确提取,首先采用基于直方图的非均衡拉伸算法对原始图像进行处理,以提高整体对比度;然后采用基于小波分析的能量特征图像法区分目标区域和背景区域,这是因为能量特征的选取与背景纹理的方向无关,对目标的变化具有旋转、平移不变性,当目标区域越平滑,检测的精度就越高。对比实验验证了该方法对起伏背景下人造目标具有较强的自适应性和良好的分割效果。The technology of segmentation based on transition region extraction was implied. To realize the extract extraction of transition region, first, a method of un-balanced tension was proposed to adjust contrast, which based on histogram. And then, energy feature image on wavelet analysis was used to differentiate the region of background and target, because the selection of energy feature was foreign to direction of background＇ s texture, target＇ s change had translation invariability, and more smooth more precise, Finally, contrast experiments demonstrate that this method leads good adaptability and super segmentation effect.

关 键 词：红外图像 过渡区 能量特征 分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]