摘要
为了解决当前基于加权平均机制的遥感图像融合算法忽略了图像像素间的差异性,易导致融合图像出现振铃及光谱失真等问题,设计了非下采样Shearlet变换耦合差异度量的遥感图像融合算法。引入亮度-色调-饱和度(IHS)变换,获取多光(multi-spectral,MS)图像中的亮度I成分;借助非下采样Shearlet变换,计算出I分量及全色(PAN)图像对应的低频及高频系数;再利用结构相似度(SSIM)函数,对低频系数的差异性实施度量,从而构建了差异度量法则,根据低频系数间的差异性采取不同的低频系数融合策略,实现低频系数的融合;考虑图像的平均梯度信息,构建高频系数融合机制,完成高频系数的融合。基于非下采样Shearlet逆变换,对融合的高频及低频系数实施处理,获取融合结果。实验结果显示,较已有的遥感图像融合方法而言,所提技术具备更高的融合效果,其输出结果含有更大的信息熵值和更低的光谱偏差度。
In order to solve the problem that the current remote sensing image fusion algorithm based on weighted averaging mechanism neglects the difference between image pixels and easily leads to ringing and spectral distortion of the fused image, this paper designs a remote sensing image fusion algorithm based on nonsubsampled shearlet transform coupling difference measure rule. IHS transform is introduced to obtain the brightness(I) components in multispectral images, and the low-frequency and high-frequency coefficients of I components and panchromatic images are calculated by nonsubsampled shearlet transform. By using structural similarity function, the difference of low frequency coefficients is measured, and the difference measurement rule is constructed. According to the difference of low frequency coefficients, different fusion strategies of low frequency coefficients are adopted to achieve the fusion of low frequency coefficients. Based on the average gradient information of the image, the fusion mechanism of high frequency coefficients is constructed to complete the fusion of high frequency coefficients. Based on the nonsubsampled shearlet inverse transform, the fusion coefficients of high and low frequencies are processed to obtain the fusion results. The experimental results show that, compared with the existing remote sensing image fusion methods, the proposed technology has higher fusion effect, and its output results contain larger information entropy and lower spectral deviation.
作者
崔怡文
任佳佳
Cui Yiwen;Ren Jiajia(Wuhan Railway Vocational College of Technology,Wuhan 430000,China;Jiangsu University of Science and Technology,Zhenjiang 212003,China)
出处
《电子测量与仪器学报》
CSCD
北大核心
2019年第12期147-154,共8页
Journal of Electronic Measurement and Instrumentation
基金
国家自然科学基金(71331003)
湖北省自然科学基金(2018CFB586)
江苏省高等学校自然科学研究面上项目(19KJB4600111)
湖北省创新行动计划(2019)资助项目。
