基于提升小波的古铜镜X光图像融合方法研究

Study on fusion method of X-ray images of ancient bronze mirrors based on lifting wavelet

为了将古铜镜X光图像信息综合到同一图像中,采用提升小波的方式将源图像进行分解,并分别对低频、高频采用不同的融合规则进行了图像融合。低频采用区域能量与区域方差相结合的方法,高频采用空间频率加邻域像素点规范中间像素点的方法,最后经提升小波逆变换得到目标图像;同时进行了理论分析和实验验证,取得了融合图像的信息熵、平均梯度和标准差数据。结果表明,在3组实验中,相对于其它几种算法,本文中算法的信息熵平均提升了5.76%,平均梯度平均提升了28.70%,标准差平均提升了7.70%,算法有效地保留了源图像的信息,对于边缘的传递效果更优秀。这一结果对古铜镜X光图像的融合是有帮助的。

In order to synthesize X-ray image information of bronze mirrors into the same image,source image was decomposed by lifting wavelet.Different fusion rules were used for image fusion at low and high frequencies respectively.The method of combining regional energy with regional variance was used in low frequency.Spatial frequency combined with the method of neighborhood pixels to standardize intermediate pixels was used in high frequency.Finally,the target image was obtained by lifting inverse wavelet transform.Theoretical analysis and experimental verification were carried out.Information entropy,average gradient and standard deviation of the fused image were obtained.The results show that,in three groups of experiments,compared with the other three algorithms,information entropy of the proposed algorithm in this paper is increased by 5.76%on average.Average gradient is increased by 28.70%.Standard deviation is increased by 7.70%on average.The algorithm effectively preserves the information of the source image.Edge transmission effect is better.This result is helpful for the fusion of X-ray images of bronze mirrors.