期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2004
卷号：XXXV Part B2
页码：543-547
出版社：Copernicus Publications
摘要：This paper is concerned with a comparative study and implementation of different image correlation techniques for Automatic Inner Orientation in aerial images. The implemented image correlation techniques are: Cross Correlation Function (CCF), Binary Cross Correlation Function (BCCF) and Least Square Matching (LSM). The first two approaches are used to determine the approximate position of the fiducial mark centres, which is then followed by a quadratic surface fitting for precise fiducial centre determination. Wiener and constrained least square filter were used as pre-processing techniques to improve the accuracy of algorithms. All three algorithms are applied to the digital images with different resolution (namely: 15, 30 and 60 micrometers pixel size). The test results show that the optimum solution can be achieved by a combination of CCF for approximate positioning followed by LSM method, i.e. LSM approach shows better performance as far as pointing precision for fiducial mark(cross type)centres have been 4.2 and 4.7 micrometers respectively when the optimum solution (i.e. CCF plus LSM)is used. Other aspects of fully automatic Digital Inner Orientation with respect to image frame direction, and determination of whether or not a mirror image is scanned are also investigated. The designed algorithms have the capability to achieve the accuracy mentioned above in any frame from different aerial cameras
