本研究提出DLBP法，它使用全偏極合成孔徑雷達（fully polarimetric SAR）資料來進行目標物分類處理。儘管地表物和雷達波交互的物理輻 射及幾何性質變化非常複雜，但是全偏極協變方矩陣（polarimetric covariance matrix）提供了偏極資訊以及偏極間之交互作用項，完整保有地物散射特性指標。
　　由於合成孔徑雷達資料含有斑駁（speckle）雜訊，會導致分類正確率降低。所以，應用全偏極合成孔徑雷達資料於目標物分類之前，預先施行斑駁抑減（speckle reduction）處理（例如多觀點平均處理），以降低雜訊對分類的影響。由於受到解析力的限制，因此在一個像元內常發生混雜著其他類別的雷達散射資訊之情形，在文中將模糊分類的概念引入於目標物分類中，使得目標物分類能夠容許類別混雜的情形存在。
　　有鑑於傳統的Euclidean距離並不符合全偏極雷達資料特性，因此本研究除了使用傳統Euclidean距離之外，並採用複數Wishart統計距離，於分類時使用。另外，由於目標物分類的目的在於判釋地物，不在於反衍，因此針對資料特性進行尺度調整預處理，以提供較高可分性以及較佳SNR的極化資料，提高類別之間的可分性。
　　結合BP類神經網路及fuzzy c-means演算法，倒傳遞類神經網路可加入額外的輔助資訊，fuzzy 理論可以提高對混合像元處理的能力，結合兩者的優點，合併成為ㄧ監督式之DLBP分類器。
　　實驗顯示可分性比SNR更適合作為尺度調整的指標，與只經過雜訊抑制的處理後的可分性相比，尺度調整後類別之間的可分性均有提升，提高1.01~1.34倍。最短距離法的整體分類精度89.93%~96.41%，DLBP法的整體分類精度89.37%~94.40%。從地真資料裡剔除不良的道路類別後，於Euclidean空間中DLBP的使用者分類精度提升約1%~5%。

　　Because fully polarimetric SAR (POLSAR) provides more information about the scattering characteristics of the earth’s surface, it enables a more accurate classification than single-channel and single-polarization SAR do. POLSAR gives all polarimetric properties in the covariance matrix. Due to speckle noise of POLSAR data, the fine pattern is difficult to be identified and the true value of back-scattering is contaminated by noise. Hence, the speckle noise filter approach is utilized to reduce noise disturbance. Moreover, the fuzzy approach is utilized to take mixed pixels or regions into account. We combine fuzzy c-means algorithm with back-propagation neural network to design a supervised dynamic learning back-propagation (DLBP) classifier.
　　A distance measure based on the complex Gaussian distribution was applied to represent the distance between any two classes in the feature space. After speckle noise filtering, we propose a method to change the data scale. The new method could improve the signal quality (higher SNR) and provide data with better separability between two classes.
　　Test results show that separability measure might be better than SNR for defining the data scale for target classification. After changing the data scale, separability measures could raise 1.01~1.34 times than Lee filtered data. Accuracy of DLBP image classification is 89.37~94.40% . To eliminate the road class whose ground truth is incorrect, the user’s accuracy increased about 1%~5%.

[1].阮勝裕、鄭克聲，1994，「類神經網路應用於地表覆蓋分類之研究」，第十三屆測量學 術及應用研討會，pp.221-229。
[2].林佳駒，2003，「特徵為基礎的模糊演算法進行多波段-多極化NASA/JPL POLSAR影像分 類之研究」，國立成功大學測量工程所碩士論文。
[3].邵泰璋、史天元，2000，「類神經網路於多光譜影像之應用」，航測及遙測學刊，5(1):1-14。
[4].高文亮、陳錕山、顏世坤、曾裕強、陳滌清，1993，「偏極化SAR影像最佳分辨器之發展」，第十二屆測量學術及應用研討會，pp.349-392。
[5].陳家堂，2002，「全偏極合成孔徑雷達於目標分類之研究」，國立中央大學太空科學研究所博士論文。
[6].陳繼藩、陳錕山、曾裕強、倪誠隆、高文亮，1993，「類神經網路應用於SPOT衛星影像土地分類之研究」，第十二屆測量學術及應用研討會，pp.331-344。
[7].陳繼藩、陳世旺、徐守道，1995，「非監督性類神經網路應用於SPOT衛星影像最佳化分類之研究」，第十四屆測量學術及應用研討會，pp.285-294。
[8].黃微萍、蔡代華、陳錕山、曾裕強，1994，「應用多頻多偏極SAR影像於農作物之分類」，第十三屆測量學術及應用研討會，pp.339-348。
[9].張維哲，1992，「人工神經網路」，全欣資訊圖書出版。
[10].鄧敏松，1996，「結合多時段遙測影像，耕地坵塊與領域知識之區域式影像辨識法於水稻田耕作調查之應用」，國立成功大學測量工程學系碩士論文。
[11].蔡敏之，1996，「應用輔助資訊及類神經網路於遙測影像土地利用分類之研究--以識別都市區為例」，國立成功大學測量工程學系碩士論文。
[12].蘇木春、張孝德，2003，「機器學習：類神經網路、模糊系統以及基因演算法則」，全華科技圖書出版
[13].Azimi-Sadjadi, M.R., Ghaloum, S. and Zoughi, R., 1993, “ Terrain classification in SAR images using principal components analysis and neural networks”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 31, Issue: 2 , pp. 511 -515.
[14].Bárdossy, A. and Samaniego, L., “Fuzzy rule-based classification of remotely sensed imagery”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 40, No. 2, pp. 362-374, 2002.
[15].Benz, U., and Pottier, E., 2001, “Object based analysis of polarimetric SAR data in alpha-entropy-anisotropy decomposition using fuzzy classification by eCognition”, Geoscience and Remote Sensing Symposium, 2001. IGARSS '01. IEEE International , Vol. 3 , pp. 1427 -1429.
[16].Bischof, H., Schneider, W. and Pinz, A.J. , 1992, “Multispectral classification of Landsat-images using neural networks”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 30 , Issue: 3 , pp.482 – 490.
[17].Boerner, Wolfgang-Martin, 2003, “WB6 interferometric and Polarimetric SAR Ⅱ, Paper WB6-5”, ACRS 2003 ISRS, Busan, KOREA.
[18].Chen, Chia-Tang, Chen, Kun-Shan and Lee, Jong-Sen, 2003, “The Use of Fully Polarimetric Information for the Fuzzy Neural Classification of SAR Images”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 41, pp. 2089–2100.
[19].Civio, D.L., 1993, “Artificial neural networks for land-cover classification and mapping”, International Journal of Geographical Information System, Vol. 7, No. 2, pp.173-186
[20].GlobeSAR-2 Program, 2001, “Educational Resources for Radar Remote Sensing”, Canada Centre for Remote Sensing, Canada. Downloaded on 2004/07/05 from the website
http://www.ccrs.nrcan.gc.ca/ccrs/rd/programs/globsar/gsar2pub_e.html
[21].Gonzalez, R., C., Woods, R., E., 1992, “Digital Image Processing”, Addison-Wesley, USA.
[22].Hara, Y., Atkins, R.G., Yueh, S.H., Shin, R.T. and Kong, J.A., 1994, “ Application of neural networks to radar image classification”, IEEE Transactions on Geoscience and Remote Sensing, Vol.32, Issue: 1, pp.100-109.
[23].Heermann, P.D. and Khazenie, N., 1992, “Classification of multispectral remote sensing data using a back-propagation neural network”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 30, Issue: 1, pp. 81 – 88.
[24].Heermann, P.D. and Khazenie, N., 1990 ,“Analysis of large multi-dimensional data with a backpropagation neural network” , IJCNN International Joint Conference on Neural Networks, Vol.1, pp. 525 – 530.
[25].Hellmann, M. and Jager, Gunther, 2002, “Fuzzy rule based classification of polarimetric SAR data”, Aerospace Science and Technology, Vol.6, Issue: 3, pp. 217-232.
[26].Hellmann, M., Jager, G., Kratzschmar, E. and Habermeyer, M., 1999, “Classification of full polarimetric SAR-data using artificial neural networks and fuzzy algorithms”, Geoscience and Remote Sensing Symposium, IGARSS '99 Proceedings. IEEE International, Vol. 4, pp. 1995 -1997.
[27].Henderson , Floyd M. and Lewis, Anthony J., 1998 , “Principles & Application of IMAGING RADAR”, J. Wiley Press, New York.
[28].Hopfield, J. , 1982, “Neural Networks and Physical System with Emergent Collective Computational abilities”, Proceedings of the National Academy of Sciences, Vol. 74, pp. 2554-2558.
[29].Huang, S.-C. and Huang, Y.-F., 1991, “Bounds on the number of hidden neurons in multilayer perceptrons”, IEEE Transactions on Neural Networks, Vol. 2 , Issue: 1, pp. 47 – 55.
[30].Kähny, D., Schmitt, K. and Wiesbeck, W., 1992, “Calibration of bistatic polarimetric radar systems”, IEEE Transactions on Geoscience and Remote Sensing, Vol.30, No. 5, pp. 847-852.
[31].Kozlov, A. I., Ligthart, L. P. and Logvin, A. I., 2001, “Mathematical and physical modelling of microwave scattering and polarimetric remote sensing :/monitoring the Earth's environment using polarimetric radar : formulation and potential applications”, Kluwer Academic Press., Netherlands
[32].Kostinski, A.B. and Boerner, W.M., 1987, “On the polarimetric contrast optimization”, IEEE Transactions on Antennas and Propagation, Vol. 35, Issue: 8, pp. 988 – 991.
[33].Kulkarni, Arun D., 1994,“Artificial neural networks for image understanding”, Van Nostrand Reinhold Press., USA
[34].Kung, S.Y. and Hwang, J.N., 1988 , “ An algebraic projection analysis for optimal hidden units size and learning rates in back-propagation learning”, IEEE International Conference on Neural Networks, Vol. 1 , pp. 363 – 370.
[35].Kung, S.Y., 1993, “Digital neural networks”, Prentice Hall press. , Canada
[36].Lee, J. S. and Hoppel, K., 1992, “Principal componets transformation of multifrequency polarimetric SAR imagery”, IEEE Trans. Geosci. Remote Sensing, Vol. 30, No. 4, pp.686 – 696
[37].Lee, J. S. and Grunes, M. R. , 1994, “Intensity and phase statistics of multilook polarimetric and interferometric SAR imagery”, IEEE Trans. Geosci. Remote Sensing, Vol. 32, No.5 , pp. 1017-1028.
[38].Lee, J. S., Grunes, M.R. and Grandi, G., 1999, “Polarimetric SAR speckle filtering and its implication for classification”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 37, pp. 2363–2373.
[39].Lee, Ken Yoong, Liew, Soo Chin, Kwoh, Leong Keong and Nakayama, Mikiyasu, 2001, “LAND COVER CLASSIFICATION USING NASA / JPL POLARIMERIC SYNTHETIC APERTURE RADAR (POLSAR) DATA”, Proc. ACRS 2001 – 22nd Asian Conference on Remote Sensing, 5-9 November 2001, Singapore. Vol. 2, pp. 1049-1054.
[40].Lillesand T. M. and R. W. Kiefer, 2000, “Remote Sensing and Image Interpretation”, John Wiley & Sons, New York.
[41].Looney, Carl G., 1997, “Pattern recognition using neural networks :theory and algorithms for engineers and scientists”, Oxford University Press., USA
[42].Masters, Timothy., 1993, “Practical neural network recipes in C++”, Academic Press., USA.
[43].NASA JPL, 2003, Technical Manuals： AIRSAR General Reference Manual. http://airsar.jpl.nasa.gov/documents/instrument.htm
[44].Rignot, E., Chellappa, R. and Dubois, P., 1992, “Unsupervised segmentation of polarimetric SAR data using the covariance matrix”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 30 Issue: 4, pp. 697 -705.
[45].Rosenblatt, F., 1958, “The perceptron : A probabilistic model for information storage and organization in the brain”, Psychological Review, Vol. 65, pp. 386-408.
[46].Rumelhart, D.E., Hinton, G.E. and Williams, R.J., 1986, “Learning internal representations by error propagation”, Parallel Distributed Processing, Vol. 1, Chap. 8, MIT Press., USA
[47].Schürmann, Jürgen., 1996, “Pattern classification :a unified view of statistical and neural approaches”, Wiley Press., New York.
[48].Tzeng, Y. C., Chen, K. S., Kao, W. L. and Fung, A. K., 1994, “A Dynamic Learning Neural Network for Remote Sensing Application”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 32, No. 5, pp. 1096-1102.
[49].Van Zyl, J.J. and Ulaby, F.T., 1990, “Scattering matrix representations for simple targets”, Radar polarimetry for Geoscience Applications, Artech House, Norwood, MA.
[50].Wiley, Carl C.A., 1965, “Pulsed Doppler Radar Methods and Apparatus”, United States Patent, No. 3, 196, 436, Filed August 1954
[51]. Yuyu Zhou, Hong Chen and Qijiang Zhu, 2002, ”The research of classification algorithm based on fuzzy clustering and neural network”, Geoscience and Remote Sensing Symposium, 2002. IGARSS '02. 2002 IEEE International , Vol. 4 , pp. 2525 -2527.