vised classification）與非監督性分類（unsupervised classifi-
cation）。監督性分類在訓練階段通常需要人工介入選取訓練區，當
類別較混雜時，此階段將是件費時且費力的工作。非監督性分類通
常需要人工介入部分是類別數目的給定，相對而言則較自動化。但
是，如何給定正確的類別數目這也是非監督性分類最困難的部分，
依據前人研究得知，可利用分類指標（cluster validity index）
作為類別數目判斷的參考。
一般非監督性的分類指標可分明確分類（Hard c-Means）指標
與模糊分類（Fuzzy c-Means）指標。台灣地區由於土地高度開發，
類別混淆情形相當常見，以明確分類方法（如ISODATA）為基礎的
指標容易判斷錯誤，而採用模糊分類方法為基礎的指標因可利用歸
屬值（membership）與模糊程度指標（fuzziness index）調整，相
對來說較具優勢。
一些相關研究顯示，分類指標必須同時考慮類別本身緊密性
(compactness)與類別之間分離性（separation）。因此，本研究嘗
試利用緊密性與分離性的模糊分類指標，應用於台灣地區多光譜衛
星影像非監督性分類的類別數目自動判斷。

to the mixture of the land cover clusters is normally found in the remote sensing image, it is also important to assign the proper fuzziness index for the fuzzy objective function before the fuzzy analysis can be performed.
This study employs a fuzzy clustering validity function, which
produces a validity index by calculating the overall average
compactness and separation of fuzzy clusters, to measure the valid number of the clusters and estimate the applicable fuzziness index for remote sensing image. Since the validity function is designed to detect compact and separated clusters, the smallest validity index certainly indicates an optimized measurement for cluster numbers and fuzziness index.
In order to obtain the smallest validity index, this study designs a series of measurements to compute the validity indexes against different fuzziness indexes and various cluster numbers. The measurements are implemented by using simulated multi-spectral data and SPOT images. The simulated image is used to obtain the applicable range of the fuzziness index. The preliminary result indicates that the fuzziness indexes ranging from 2 to 2.5 are able to cope with the relatively fuzzy images. The primary testing of SPOT images demonstrates that the fuzzy clustering validity function is able
to provide the proper guide for the determination of the number of clusters when the FCM algorithm is applied to satellite image classification.

王玲玲，周紀夢著，常用統計方法，華東師範大學出版社，民國83年
徐守道，應用非監督性類神經網路於SPOT衛星影像分類最佳化之研究，碩士論文，國立中央大學土木工程研究所，中壢，1995
曾國雄，多變量解析與其應用，華泰書局，民國74年
藎壚，實用模糊數學，亞東書局，民國80年
Bezdek, J. C., 1973. Fuzzy mathematics in pattern classification,
Ph. D. Dissertation, Appl. Math. Cornell University, Ithaca,
New York.
Bezdek, J. C., 1980. A convergence theorem for the Fuzzy ISODATA clustering algorithms. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. PAMI-2, No.1, p.1-8.
Bezdek, J. C., 1981. Pattern recognition with fuzzy objective function algorithms, Plenum Press, New York.
Bezdek, J. C. and Pal, Nikhil R., 1998. Some new indexes of cluster validity. IEEE Transactions on Systems, Man and Cybernetic-Part B: Cybernetics, Vol.28, No.3, p.301-315.
Boudraa, A. -O., 1999, Dynamic estimation of number of clusters in data sets. Electronics Letters, Vol.35, No.19, p.1606-1608.
Davies, D. L. and Bouldin, D. W., 1979.A cluster separation measure. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. PAMI-1, No.2, p.224-227.
Duda, R. O. and Hart, P. E., 1973: Pattern classification and scene analysis, Wiley Interscience, New York.
Dunn, J. C., 1973.A fuzzy relative of the ISODATA process and its use in detecting compact well-separated clusters. Journal of Cybernetics, 3, 3, Jul-Sept., p.32-57.
Dunn, J. C., 1974.Well-separated clusters and optimal fuzzy partitions. Journal of Cybernetics, 4, 1, p.95-104.
Dunn, J. C., 1974. Some recent investigations of a new fuzzy partitioning algorithm and its application to pattern classification problems. Journal of Cybernetics, 4, 2, p.1-15.
Gunderson, M., 1978. Application of fuzzy ISODATA algorithms to star tracker pointing systems. Proceedings of 7th Triennial world IFAC Cong., Helsinki, Filind, p.1319-1323.
Hassar, H. and Bensaid, A. , “Validation of fuzzy and crisp c-partitions”, Fuzzy Information Processing Society, 1999. NAFIPS. 18th International Conference of the North American , p.342-346 , 1999
Kwon, S. H., 1998.Cluster validity index for fuzzy clustering. Electronics Letters, Vol.34, No.22, p.2176-2177.
Pal, N. R. and Bezdek, J. C., 1995.On cluster validity for the fuzzy c-means model. IEEE Transactions on Fuzzy System, Vol.3, No.3, p.370-379.
Pal, N. R. and Bezdek, J. C., 1997.Correction to On cluster validity for the fuzzy c-means model. IEEE Transactions on Fuzzy System, Vol.5, No.1, p.152-153.
Rhee, H. -S. and Oh, K. -W., 1996.A validity measure for fuzzy clustering and its use in selecting optimal number of clusters. Proceedings of the Fifth IEEE International Conference on Fuzzy Systems, Vol.2, p.1020-1025.
Rhee, H. -S. and Oh, K. -W., 1996.A performance measure for the fuzzy cluster validity. Soft Computing in Intelligent Systems and Information Processing. Proceedings of the 1996 Asian Fuzzy Systems Symposium, p.264-369.
Rezaee, M. R., Lelieveldt, B. P. F. and Reiber, J. H. C., 1998.A new cluster validity index for the fuzzy c-mean. Pattern Recognition Letters, Vol.19 p.237-246.
Tou, J. T. and Gonzalez, R. C., 1974.Pattern Recognition Principles.Addison-Wesley.
Windham, M. P. 1982.Cluster validity for the fuzzy c-means clustering algorithm. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. PAMI-4, N0.4, p.357-363.
Xie, X. L. and Beni, G., 1991.A validity measure for fuzzy clustering. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.13, No.8, p.841-847.
Zahid, N., Limouri, M.and Essaid, A., 1999.A new cluster validity for fuzzy clustering. Pattern Recognition, Vol.32, p.1089-1097.
Zimmermann, Hans J., 1984.Fuzzy set theory and its applications: Kluwer-Nijhoff publishing.