摘要
在非多尺度分解的框架下,针对像素级的图像融合,提出了一种基于统计信号处理图像融合算法。该算法从一个直观的图像融合模型出发,即所获得的原始图像是模糊因子乘理想的融合图像与噪声进行非线性累加所得,应用模糊数学的原理和统计信号处理的EM算法,递推求出噪声的各项参数和对应的模糊因子,最终导出所求的融合图像。试验表明,应用该算法融合后的图像取得了满意的效果。
In the frame of non- multi- scale decomposition, a new fusion algorithm based on statistical signal processing and fuzzy mathematics is proposed for pixel - level picture fusion. A fusion model, which assumes that the source picture is the sum of stochastic noise and ideal fusion picture multiplied by a fuzzy factor, is considered, and its noise parameters and corresponding fuzzy factor are calculated by the iteration based on Fuzzy mathematics and EM algorithm of statistical signal processing so that the fusion picture is gotten. The experimental results show that the better image quality is obtained in the proposed algorithm.
出处
《计算机与数字工程》
2006年第7期42-44,48,共4页
Computer & Digital Engineering
关键词
统计信号处理
非多尺度分解
EM算法
图像融合
Statistical signal processing,Non- multi - scale decomposition, EM algorithm, Image fusion
