在過去的幾十年裡，影像處理的邊緣偵測問題已經被研究多年了。大部份的邊緣偵測問題都著眼在找出邊緣的位置。在本研究中，我們提出一個方法是利用『經驗解模法』偵測邊緣。在這個方法中我們會利用到關於目標物大小的資訊，並利用它來偵測線性的目標物，例如河川。

　　經驗解模法在本質上可以把資料分成一組可使用希伯特轉換的〝本質模態函數〞，這方法適用且有效率，由於它是築基於區域的特徵時間尺度，它可用之於非線性及非平穩資料的處理，經由希伯特轉換後，〝本質模態函數〞會顯出不同時間的瞬間頻率，最後產生一個〝能量─頻率─時間〞的分佈，名之為〝希伯特頻譜〞，這方法的主要觀念是建基於訊號的區域特性，而使瞬間頻率有它的意義，如此就不需像古老方法那樣硬要把非線性及非平穩的訊號以調和方程表示。我們可以在物體的邊緣上發現高的瞬間頻率，這個發現導引我們一個偵測邊緣的方向。
　
　　在本研究的最後利用希伯特頻譜去增加高維度的特徵。我們在將頻率軸切細分成幾等分，並紀錄其對應的振幅值，如此會產生一組由高到低的頻譜影像。將這些這些 頻譜影像加到欲分析的光譜資料中去做分類，會發現它提升了原本的準確率。

　Edge detection has been studied substantially in image processing for the past few decades. Most methods of edge detection are focused on determining the locations of edges. In this study, we propose an approach to edge detection based on the empirical mode decomposition (EMD). Using this information about the object size, we may recognize the linear shaped objects, such as rivers.

　The EMD and Hilbert-Huang Transform (HHT) have been recently introduced in signal processing for non-linear and non-stationary signal analysis. In essence, the EMD can be used to generate a unique set of intrinsic mode functions (IMFs) for a given signal waveform. Based on the local characteristic of the waveform, instantaneous frequencies at each point can be derived from the IMFs. We have found
that high instantaneous frequencies occur at the edges of objects coincidentally. This finding gives us a guideline for edge detection.

　Another method proposed in this study to increase the dimensions of the features based on the HHT. After analyzing the frequency domain computed by the HHT, we can divide the frequencies into several groups, ranging from high to low frequencies. These different frequency groups are like the frequency analyzed by wavelet.

　The proposed approach is applied to real multispectral images. Our preliminary results show that the proposed approach can successfully detect edges in synthetic images and rivers in real images.

[1] N. E. Huang, Z. Shen, S. R. Long, M. C. Wu, H. H. Shih, Q. Zheng, N. Yen, C. C.
Tung and H. H Liu, “The Empirical Mode Decomposition and the Hilbert
Spectrum for Nonlinear and Non-Stationary Time Series Analysis, Proceedings,’’
in Proc. R. Soc. London A, vol. 454, pp. 903–995, 1998.
[2] A. Linderhed, “2-D empirical mode decompositions – in the spirit of image
compression,’’ in Wavelet and Independent Component Analysis Applications IXI,
vol. 4738, pp.1–8, April 2002.
[3] P. Flandrin, G. Rilling, and P. Goncalves, “Empirical Mode Decomposition as a
filter bank,’’ IEEE Signal Processing Letters, vol. 11, pp. 112–114, 2004.
[4] G. Rilling, P. Flandrin and P. Goncalvs, “On Empirical Mode Decomposition
and Its Algorithms,” Laboratoire de Physique (UMR CNRS 5672).
[5] G. Qiang, M. Xiaojiang, Z. Haiyong and Z. Yankun, “Processing time-varing
signals by a new method, ” in Proc. Radar. in Int. Con., pp.1011–1014, Oct.
2001.
[6] K. T. Coughlin and K. K. Tung, “11-Year solar cycle in the stratosphere extracted
by the empirical mode decomposition method,’’ Advances in Space Research, vol.
34, pp.323–329, 2004.
[7] J. C. Nunes, Y. Bouaoune, E. Delechelle, O. Niang, and P. Bunel, “Image
analysis by bidimensional empirical mode decomposition,’’ Image and Vision
Computing, vol.21, pp. 1019–1026, 2003.
[8] S. Beucher, “Geodesic reconstruction, saddle zones and hierarchical
segmentation,’’ Image Anal. Stereol, vol. 20, pp. 137–141, 2001.
[9] J. C. Carr, W. R. Fright, and R. K. Beatson, “Surface interpolation with radial
basis functions for medical imaging,’’ Comput. Graph. Proc., Annu., pp. 67–76,
2001.
[10] Z. X. Liu, H. J. Wang, and S. L. Peng, “Texture classification through directional
empirical mode decomposition,” Int. Conf. Pattern Recognition, vol. 4, pp. 803-
806, Aug. 2004.
50
[11] J. P. Havlick, D. S. Harding, and A. C. Bovik, “The multicomponent AM-FM
image representation,” IEEE Trans. Image Processing, vol. 5, no. 6, pp.
1094–1100, Jun. 1996.
[12] Zhongxuan Liu, Peng, S., “Boundary Processing of bidimensional EMD using
texture synthesis,’’ IEEE Signal Processing Letters, vol. 12, pp. 33–36, Jan.
2005.
[13] B. Julesz, “Textons, the elements of texture perception, and their interactions,”
Nature, vol. 290, no. 12, pp. 91–97, Mar. 1981.
[14] A. A. Efros and T. K. Leung, “Texture synthesis by nonparametric sampling,” in
IEEE Computer Vision, pp. 1033–1038, 1999.
[15] J. Canny, “A Computational Approach to Edge Detection,’’ IEEE Transactions
on Pattern
Analysis and Machine Intelligence, vol. 8, no. 6, Nov. 1986.
[16] G. M. Livadas, A.G. Constantinides, “Image edge detection and segmentation
based on the Hilbert transform,’’ in Proc. IEEE Int. Conf. Acoustics, Speech, and
Signal Processing, vol.2, pp.1152 – 1155, April 1988.
[17] M. Farge “Wavelet transforms and their applications to turbulence,’’ in Annu.
Rev Fluid Mech, vol. 24, pp. 395–457, 1992.
[18] N. E. Huang, Z. Shen, S. R. Long. “A new view of nonlinear water waves: the
51
Hilbert spectrum,’’ in Annu Rev
Fluid Mech, vol. 31, pp. 417–57, 1999.
[19] H.C. North and Q.X. Wu, “Multiscale edge detection and thresholding in SAR
imagery,” in Proc. Image and Vision Computing, pp. 247–252, Aug. 1999.