AMOEBA影像分類技術之研究

Abstract

本研究探討一種結合光譜資訊和空間資訊的分類法，稱為AMOEBA，並與 ISOCLASS、K-means等常用的群聚法做比較，以期瞭解加入空間資訊後， 影像分類的情形。另外也探討AMOEBA、ISOCLASS、K-means等群聚法，所 需的參數與參數設定的難易程度。 This study investigates a clustering method which incooperates both spectral and spatial information, named AMOEBA. The cation results are compared with ISODATA and K-means. The setting issues of the parameters for AMOEBA, ISODATA and K- means are also addressed.