Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/2289
標題: 影像處理結合人工智慧於相位重建法之研究
作者: 賴乾志
Lai, Cian-Jhih
關鍵字: phase unwrapping;相位重建;fuzzy logic;image processing;模糊邏輯;影像處理
出版社: 機械工程學系
摘要: 
本論文主要目的是對現有的相位重建技術,及其相關之基礎學術與應用技術做一深入的探討研究,並針對這些方法之缺點進行改善,擬建立一快速且無誤差的相位重建技術。
論文中首先針對平行性相位重建法(Parallel Phase Unwrapping)的特性以及性能進行探討,進而提出一個較為簡潔、快速且強健的平行性相位重建法,此方法可直接對高奇異點密度分佈之相位圖進行重建,無需先對奇異點進行偵測及處理。
接著本論文更提出一利用phase jumps分區塊的相位重建法,此方法以全域(global)及區域(local)混合處理之方式,進行高效能之相位重建。首先對無phase jump之區域執行全域的相位重建,隨後處理phase jump區的相位重建;它大幅地節省了全域處理區所須之相位重建時間,因此此法擁有相當快速的相位重建速度。
最後結合人工智慧之模糊(Fuzzy)理論的應用,以決定各點之偏差(Bias)值直接進行相位之重建,其速度與精度皆遠較前述之方法為佳。

An innovative phase unwrapping algorithm is proposed here. It has a parallel and noise-immune nature. The criterion itself is simple, concise and robust. Unlike earlier parallel unwrapping algorithm, which converts the wrapped data into a saw-tooth format as soon as possible, the proposed algorithm converts the wrapped map into a clear map in several times of iteration and then shifts the clear fringes, one pixel per frame, outside of the map.
Advantageously, the algorithm could be applied only to the data, located on the wrapped band, leaving data elsewhere with an appropriate global movement. In this approach, null dummy work is done during the hybrid implementation. Therefore, it is a highly efficient work. Finally, a fuzzy logic is applied to the wrapped data to decide the bias of each pixel of the map. Once done, the unwrapped phase field could be successfully achieved.
URI: http://hdl.handle.net/11455/2289
Appears in Collections:機械工程學系所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.