| Summary: | 碩士 === 國立中興大學 === 土木工程學系 === 87 === Image Matching theory of conjugate points on digitial aerial images has been developed since 1950’s,and it tends to be matured today. But there are often faultly matching happening in practical applications. This is because of poor image quality,uniform tones of matching areas and casting shadow. Especially when casting shadow exists,it varies large gray value of image and causes extra diffculty of matching conjugate points on shadow areas. Even though we apply image matching on shadow areas,we are not necessarily get accurate conjugate points. And that is to say nothing of the ability to match accurate conjugate points by automatic Image Matching method.
The objectives of this paper are to detect the shadow regions of aerial images,to restore gray value of shadow regions in no shadow-casting conditions,and simultaneously enhance the accuracy of matching. First,we choose shadow areas by manual operations,and then find exact shadow regions in these areas by applying to Image Segmentation method cooperating with Mathematical Morphology theory. Second,we adjust image histograms of shadow regions and restore their gray value in no shadow-casiting conditions by Image Histogram Matching method. Finally,we apply image matching of conjugate points in shadow regions of two stereo overlapping imges respectively before and after shadow restoration processes. That will assist us to examine the ability of enhancing accuracy of image matching in aerial shadow areas,and it can also demostrate the objectives of this paper.
|
|---|
