二维Hilbert-Huang变换及其在图像增强中的应用

Bidimensional Hilbert-Huang transform and its application in image enhancement

为了更加有效地提取图像细节,在分析希尔伯特——黄变换(Hilbert-Huang Transform,HHT)的基础上给出了二维HHT的实现方法,并应用于图像增强.首先对二维图像信号进行基于Delaunay三角分割的二维经验模式分解,再将分解得到信号的各个内蕴含模式分量分别作总体Hilbert变换.实验结果表明,此方法可细致地描绘出图像的边缘信息,并可在不同程度上体现图像的轮廓信息.该研究在图像压缩和图像分割中有重要的意义.

In order to exact the details of an image more accurately, based on the analysis of Hilbert-Huang transform （ HHT）, the realization of bidimensional HHT is provided and used to process the image. Delaunay triangulation is used in bidimensional empirical mode decomposition （BEMD） to get each intrinsic mode function （IMF）, and then each IMF is treated with total Hilbert spectral analysis （THSA）. Experimental result shows that this method can describe edge details more accurately, and contour information can be presented in different degrees. This study is of great of significance in the image compression and image segmentation.