摘要:
为进一步增强遥感图像的细节信息,在非下采样轮廓变换(non-subsampled contourlet transform, NSCT)的基础上,结合模糊理论,提出了一种遥感图像增强算法.首先对原始图像进行NSCT变换,得到频率域内低频系数和不同尺度不同子带上的高频系数;然后定义隶属度函数,对高频系数进行模糊变换;在进行NSCT逆变换重构图像的过程中,逐层地将高频系数线性地加到低频系数中,最终实现遥感图像增强.实验结果表明,该算法在主、客观方面都使遥感图像得到了很好的增强效果.研究表明,NSCT变换后的高频系数包含了图像中的细节信息,针对高频系数进行模糊变换后,进行NSCT逆变换可以比较有效地增强图像.该算法存在的问题在于计算量较大以及需要调整的参数较多.
Abstract:
A remote sensing image enhancement algorithm, which is based on the non-subsampled contourlet transform (NSCT)and the fuzzy theory, was proposed in this paper.Firstly, the low pass and high pass coefficients in different sizes of the image were acquired using the NSCT transform.Then, a membership function in fuzzy theory was defined to enhance the high pass coefficients.In the process of transforming the fuzzy domain to NSCT domain and reconstructing the image, the high pass sub-bands coefficients were added into low pass sub-bands step by step and the enhancement was realized finally.The results of the experiments show that the proposed method could enhance the remote sensing image perfectly in both subjective and objective aspects.The results obtained by the authors suggest that the high-pass coefficients of the NSCT transform of the image contain most of the details of the original image, and image enhancement task could be attained by fuzzy transformation of the high-pass coefficients.However, the proposed method has the disadvantages of large computation quantity and the requirement of manual adjustment of several parameters.
下载: