以形態單元為基底的彩色影像擷取法

Abstract

近年來，隨著電腦科技發展迅速，數位化影像資料的數量正以極快的速度增加。如何管理一個大型的數位影像資料庫也越來越重要。因此對於數位影像內容的檢索變成為一個重要的研究領域。本研究的主要目的是針對自然影像建立一個有效率的檢索方法。數學形態學具有簡單的整數運算，可平行處理，可解決影像處理上的及時問題。因此我們以數學形態學的運算擷取出影像的特徵並建立其形態單元(Morphological Primitives)，再依此計算查詢影像和資料庫中影像的相似度，選出最相近的影像。實驗結果顯示我們所提出的方法是相當快速且正確的。

In recent year, with the fast development of computer technology, the data amount of digital images is grown very rapidly. The management of the large image database becomes an important issue. So content-based image retrieval has become an important research area. The purpose of this thesis is to construct an efficient method of natural image retrieval. Morphological operators are attractive because they involve simple logical operations and can be implemented in parallel, making real-time applications possible. Thus, we utilize the advantage of morphological operators to extract the morphological primitives to represent the features in images. Then we can select the most similar images according the similarity between query image and database images. Experimental results indicate the proposed method is very fast and efficient.