合成孔徑雷達(SAR)是一種主動式的遙測系統，主要用於測定大範圍地區的地表資料。且其使用電磁波譜裡的微波波段，具有良好的大氣穿透力，適用於日夜的偵測、以及任何天氣條件下獲取資料的能力，因此，在現代遙測發展上佔有一重要地位。而在SAR影像的實務應用上，如：變遷偵測、重建方位參數、或合成孔徑雷達干涉法求定DEM…等，都需要對影像進行控制點的量測。由於雷達系統成像性質的影響，導致影像解析力不佳，且斑駁雜訊干擾嚴重，因此，以人眼來進行大量點位的量測是相當吃力的。

本研究利用成像品質較佳的航測正射影像為輔助，透過人工量測少量共軛點位的「半自動化初步套合」、或利用影像本身所提供的大地參考資料(Geo-Referencing Data)進行「全自動化初步套合」處理，先建立起影像間的初始對應關係。隨後於影像上進行特徵線的萃取，利用這些線特徵進行自動化線特徵的細匹配(fine matching)。在線的細匹配部分運用了模版匹配模式，計算模版內線特徵的重疊度，以重疊度高者為匹配位置，此法可同時考慮到局部線特徵及其周圍特徵的位相條件。並透過影像金字塔由粗到細的匹配策略，以解析力較粗一層中的套合成果作為下一層(較細一層)的起始對應關係，並且利用其套合精度作為模版搜尋範圍的約制，最後在原始影像層中得到儘可能均勻分佈於全張影像的匹配點對，以作為地面控制點之用。

實驗中所使用的影像為嘉義縣境內兩張平原地區及一張阿里山區的SAR影像，以及1/5000航空正射影像，涵蓋地面範圍約2.5×2.5(km2)及2×2(km2)，而初步成果可得到約1500或3000個(視影像條件而定)匹配點位，整體套合精度約0.8～1.4個像元。

Belonging to one kind of active remote sensing system, Synthetic Aperture Radar (SAR) is mainly used for data acquisition in a large area. By operating the microwave band in the electromagnetic spectrum, SAR is capable of penetrating the atmosphere and haze, and suitable for day and night operation in all kinds of weather conditions. Therefore, SAR is very useful in remote sensing. Today’s state-of-the-art of SAR applications are, for instance, change detection, SAR image orientation, or DEM generation by means of Interferometric SAR (InSAR), etc., in which ground control points are often needed. However, due to the properties of SAR image formation and due to SAR speckles, the SAR image resolution and quality are reduced severely and apparently. Hence, it is a very difficult and time-consuming work to measure the ground control points manually on SAR images.

Based on high quality of aerial ortho images, both semi-automatic and fully automatic registration approaches are utilized to establish the initial correspondence between SAR and aerial ortho image. In the semi-autoamtic registration approach, only as few conjugate points as possible are manually measured on both SAR and aerial ortho images. In the fully automatic approach, the known geo-referencing data of SAR images are used. Sequentially, line features are extracted from both SAR and aerial ortho images and used for automatic fine matching based on template matching. The overlap percentage of line features is calculated in each template and it is considered that the higher the percentage, the higher possibility the correctly matched location. Simultaneously, the topology between each line feature and its neighbor features is taken into account. A coarse-to-fine strategy is also adopted by applying an image pyramid, where the searching range in each image level is statistically defined by the a posteriori standard deviation of image registration on its upper level. Finally, as plenty of well distributed matched point pairs as possible are automatically determined and can be utilized in later applications such as SAR image orientation.

Tests are done using the 1:5000 aerial ortho images and the NASA/JPL TOPSAR images, which covers c.a. 2.5 km × 2.5 km of plain areas and about 2 km × 2 km of Mountain Ali area in Jia-Yi County, respectively. Test results show that the proposed methods can determine about 1500~3000 matched points in c.a. 30 minutes on a standard PC with a CPU speed of 1.4 GHz. The registration accuracy of 0.8~1.4 pixel is reached.

[1].江凱偉，1997，小波分解輔助影像紋理區之分塊，國立成功大學測量工程研究所碩士論文
[2].邱式鴻，1996，Förstner特徵物萃取法精確性的探討，第十五屆測量學術及應用研討會論文集(下冊)，pp. 667-676
[3].陳良健，李政道，1999，特徵匹配於影像起始套合應用之研究，第十八屆測量學術及應用研討會論文集(二)，pp. 683-693
[4].陳良健，2002，脈動：從數位航測到多元探測，第二十一屆測量學術暨應用研討會—科技演講，國立交通大學，新竹
[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].陳晉徹，1993，數值影像中相同紋理區分塊方法之研究，國立成功大學測量工程研究所碩士論文
[7].張奇，1995，空載合成孔徑雷達影像於外方位參數與地表重建之研究，國立中央大學土木工程研究所碩士論文
[8].湯凱佩，1995，相對方位自動化之研究，國立成功大學測量工程研究所碩士論文
[9].黃華尉，2001，TWD97與TWD67二度TM坐標轉換之研究，國立成功大學測量工程研究所碩士論文
[10].曾清涼，儲慶美，1999，GPS衛星測量原理與應用，國立成功大學衛星資訊研究中心
[11].蔡奇璋，1995，以半自動影像匹配實施與解析製圖儀結合之空中三角測量，國立成功大學測量工程研究所碩士論文
[12].蔡展榮，1997，參數型立方迴旋內插法，中國土木水利工程學刊，第九卷，第三期(民國八十六年)，pp. 483-490
[13].趙鍵哲，1988，偵測邊緣線以協助影像匹配之研究，國立成功大學測量工程研究所碩士論文
[14].叢蕙滋，1997，特徵匹配於影像套合之應用，國立中央大學土木工程學研究所碩士論文
[15].Ali, M.A., Clausi, D.A., 2002, Automatic Registration of SAR and Visible Band Remote Sensing Image, Geoscience and Remote Sensing Symposium, 2002. IGARSS '02. 2002 IEEE International, Vol. 3, pp.1331 -1333
[16].Brown, L., 1992, A Survey of Image Registration Techniques, ACM Computer Surveys, Vol. 24, No. 4, pp.325-376
[17].Curlander, J.C., and, McDonough, R.N., 1991, Synthetic Aperture Radar System and Signal Processing, John Wiley and Sons
[18].Chu, Anhua, 2001. AIRSAR Integrated Processor Documentation – Data Formats, Version 0.15, August 21, 2001.
[19].Dare, P., 2000, Automatic Registration of Multi-Source Remotely Sensed Images and Other Non-Image Spatial Information Products, 10th Australasian Remote Sensing and Photogrammetry Conference, Adelaide, Australia, CD-ROM
[20].Dare, P., Dowman, I., 2001, An Improved Model for Automatic Feature-Based Registration of SAR and SPOT Images. ISPRS Journal of Photogrammetry & Remote Sensing 56, pp.13-28.
[21].Definiens Imaging Corporation, 2001, eCognition 2.0 User Guide, Definiens Imaging Corporation
[22].Evans, A.E., 1996, A Gamma Filter for Multi-Look Synthetic Aperture Radar Images, International Symposium on Signal Processing and its Application, ISSPA, Gold Coast, Australia, pp.829-832
[23].Förstner, W., 1993, Feature Extraction in Digital Photogrammetry, Photogrammetry Record, 14(82)：pp.595-611
[24].Fonseca, M.G., Manjunath, B.S., 1996, Registration Techniques for Multisensor Remotely Sensed Imagery, Photogrammetric Engineering & Remote Sensing, Vol. 62, No. 9, pp.1049-1056
[25].Franceschetti, G., Lanari, R., 1999, Synthetic Aperture RADAR Processing, CRC, USA
[26].Garzelli, A., 2002, Wavelet-Based Fusion of Optical and SAR Image Data Over Urban Area, ISPRS Commission III, Symposium 2002, September 2002, Graz, Ausria, CD-ROM
[27].Gonzalez, R., C., Woods, R., E., 1992, Digital Image Processing, Addison-Wesley, USA
[28].Henderson, F.M., Lewis, A.J., 1998, Principles and Applications of Imaging Radar, New York：J. Wiley
[29].Inglada, J., Adragna, F., 2001, Automatic Multi-Sensor Image Registration by Edge Matching using Genetic Algorithms, Geoscience and Remote Sensing Symposium, 2001. IGARSS '01. IEEE 2001 International, Vol. 5, pp.2313 –2315
[30].Li, H., Manjunath, B.S., Mitra, S.K., 1995, A Contour-Based Approach to Multisensor Image Registration, IEEE Transaction on Image Processing, Vol. 4, No.3, pp.320-334
[31].Lillesand, T.M., Kiefer, R.W., 2000, Remote Sensing and Image Interpretation, John Wiley & Sons., USA
[32].Mallat, S., 1998, A Wavelet Tour of Signal Processing, Academic Press
[33].Mikhail, E.M., Bethel, J.S., McGlone, J.C., 2001, Introduction to Modern Photogrammetry, John Wiley & Sons., USA
[34].Schenk, T., 1999, Digital Photogrammetry, TerraScience, OH, USA
[35].Tsay, J.R., 2001, A New Approach for Automatic Determination of Edge Center Lines in a Digital Image, Asian Journal of Geoinformatics, June 2001, Vol. 1, No. 4, pp.15-22
[36].Wolf, P.R., Dewitt, B.A., 2000, Elements of Photogrammetry with Application in GIS, 3rd edition, McGraw-Hill
[37].Wu, Y., Maitre, H., 1990, A Multiresolution Approach for Registration of a SPOT Image and a SAR Image, Proceedings of International Geoscience and Remote Sensing Symposium, pp.635-638