影像拼貼應用高解析度 X-ray 生物影像

Abstract

隨著科技發展的日新月異，對於影像獲取上的技術也越來越成 熟，相對的能看的影像也越來越細微，解析度也日漸清晰。但受限 於硬體器材一次所能取像的區域有限，無法一次就將所有影像截取 完成。所以要借重數位影像的拼貼技術，將每張所取像的影像重新 拼貼接合在一起，建立成一張完整的大型影像或全景影像。 電腦能模擬出真實世界般的虛擬實境出來。因為如何建立一套 系統來建立逼真的虛擬世界，在影像處理領域中是相當熱門的研究。 本論文先使用亮度正規化(Lighting normalization)的方法來進行色差調整。此種方法可以有效的將影像跟影像間的亮度關係透 過正規化的方式將色差調整好，是種利用影像跟影像間的相對關係 來調整色差的方法。 將調整後的影像找尋重疊區域計算的關聯性（Correlation coefficient），重疊區域計算的關聯性的值越高表示相似度越高，也 就是兩張影像要對應的位置，如此一來就可以將影像拼貼在一起。

Along with the changes of technological progress , Gainging images in technology is also becoming more and more mature .Relatively, the phantom which can look is much more slight , the resolution is also clear day after day. But that is restricted in hardware equipment, because one time to be able to take the alike region to be limited, the means one time will not have possessed the image interception to complete. Therefore we must borrow renumbers the position image to image stitch the technology, to take the alike phantom each institute to spell pastes joins in together, then establishes a complete large-scale image.. The computer can simulate the hypothesized solid boundary which like the real world .How to establish the set of systems to the lifelike hypothesized world in the image process domain is the quite popular research. This developing paper progress is used “the Lighting Normalization Method” to adjust the color brightness. This method may effectively adjust on standardized way of the color the brightness relations between the images. And it’s also one kind of use image relates the color matching method relatively with the image. After we adjusted the image pursues the overlap region correlation coefficient, the overlap region computation's relatedness value which is higher means the expression similarity is higher, two images must correspond position. Finally, we can stitch the images together.