Please use this identifier to cite or link to this item:
|標題:||Unsupervised Classification for Remote Sensing Images
Hyperspectral imaging has become an emerging technique in remote sensing community, butit is still in its infancy in Taiwan due to the lack of understanding its potential in applications.Accordingly, one of its major goals is to facilitate research in hyperspectral imaging in Taiwan. Another is to present in professional societies and publish papers in peer reviewed journals. A third goal is to promote visibility of National Chung Hsing University in remote sensing community. A fourth goal is to establish working relationship between National Chung Hsing University and National Chiayi University and explore potential projects. Last but not least is to plant a seed for a possible setup for a remote sensing laboratory or center for excellence in national Chung Hsing University. With recent advances of hyperspecral remote sensing technology the utility of hyperspectral imagery covers a wide range of military and civilian applications. Of particular interest are targets with their small presence and low probability existence in data exploitation. Two major challenging issues arise in unsupervised classification. One is how to generate desired knowledge directly from the data in an unsupervised manner. The other is how to find an appropriate follow-up classifier to use the obtained unsupervised knowledge to perform supervised classification. This project presents a new approach to unsupervised classification for multispectral imagery. To address the first issue the pixel purity index (PPI) which is commonly used in hyperspeftral imaging for endmember extraction is used to find a good set of initial training samples without prior knowledge. To address the second issue the PPI-found samples are then used as training samples for a support vector machine to find a good set of training samples for a follow-up supervised classifier, Fisher's linear discriminate analysis (FLDA) which performs classification iteratively to produce final results.
高頻譜影像在遙測領域中已經成為一種先進的技術，由於在台灣缺乏潛在的應用，目前仍然是屬於初期發展的階段。因此，第一個主要的目標是促進在台灣高頻譜影像分析的研究。另一個藉此把研究成果發表在專業領域的論文在學術期刊上。 第三個目標是提升國立中興大學在遙測影像分析領域中的能見度。第四個目標是將建立國立中興大學和嘉義大學的合作關係，例如:森林族群的分類、檢測森林地區的崩塌位置等等。最後是對於國立中興大學在未來成立遙測影像技術實驗室或卓越遙測中心鋪路。由於高頻譜影像遙測技術已漸趨成熟，高頻譜影像實際應用上包括廣泛的軍事和民間應用。其中特別感興趣的是在資料中搜索屬於小而不易發現的目標。一般而言低出現機率的目標物。這類的目標常出現在農業和生態學方面。然而在非監督式的分類方法中，有兩項重要的議題，第一是在毫無背景知識的資料中如何產生訓練樣本。第二是如何尋找適當的分類器將資料做有效的分類。本研究將以像素純度指標(pixel purity index)簡稱PPI做為擷取訓練樣本的主要演算法並結合Fisher’s linear discriminate analysis (FLDA)將資料在非監督試的過程中獲得較好的分類。
|Appears in Collections:||電機工程學系所|
Show full item record
TAIR Related Article
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.