摘要
针对双色红外的图像融合特点,提出了一种基于在小波分解域中,在高频子带上采用区域能量测度融合规则,在低频子带上采用加权平均融合规则。首先,双色红外成像系统分别对中波和长波红外图像进行提升小波分解,提取出2幅图像的低频和高频信息,并采用不同的融合规则对低频和高频子带图像进行融合,然后通过提升小波反变化对各层系数进行重构,得到目标区域的融合图像。而后再通过小波软阈值化去噪,最后进行目标自动检测。通过实验对比证明,提出的融合规则更好地提高了目标的检测和识别概率。
New fusion rules are proposed according to the fusion characteristics of double-color infrared image,that is regional energy measurement fusion rule in high frequency sub-band and weighted average fusion rule in low frequency sub-band based on wavelet domain.At first,the double-color infrared imaging system decomposed medium-wave and long-wavelength infrared images,extracted the low frequency and high frequency information of two images,used different fusion rules on low frequency and high frequency sub-band image fusion,reconstructed wavelet coefficients of each layer through the change of ascending wavelet to gain the fused image of target area,and carried on a target automatic detection through the wavelet soft threshold de-noising.This comparative experiment proves that the fusion rules can improve the target detection and identification probability.
出处
《现代电子技术》
2011年第16期58-60,共3页
Modern Electronics Technique
关键词
双色红外
提升小波
图像融合
目标检测
double-color infrared
ascending wavelet
image fusion
target detection