DC 欄位 |
值 |
語言 |
DC.contributor | 遙測科技碩士學位學程 | zh_TW |
DC.creator | 戴政淳 | zh_TW |
DC.creator | Cheng-chun Tai | en_US |
dc.date.accessioned | 2012-8-29T07:39:07Z | |
dc.date.available | 2012-8-29T07:39:07Z | |
dc.date.issued | 2012 | |
dc.identifier.uri | http://ir.lib.ncu.edu.tw:88/thesis/view_etd.asp?URN=990202005 | |
dc.contributor.department | 遙測科技碩士學位學程 | zh_TW |
DC.description | 國立中央大學 | zh_TW |
DC.description | National Central University | en_US |
dc.description.abstract | 衛星影像中的地面光譜資訊可以協助研究者了解地面的真實狀況,但由於光譜資訊分析上的限制,以傳統的分類分法是很難精確並有效率地進行高光譜影像分類。
本研究應用親和力傳播理論來探討如何降低影像分類的瓶頸。該方法將成對的資料點進行相似度的運算,並交換成對點的實際資料值,直至一組典範及其對應的聚類出現。另外,分析過程中,同時處理大張影像內所有的資料點是非常耗時的,研究者必須思考如何提高效率,而本研究是透過結合階層式分析途徑來佳化其結果。
本研究使用階層式親和力傳播方法分析ROSIS機載光譜儀影像。研究結果顯示階層式親和力傳播分類方法相較於傳統的親和力傳播分類方法,能夠降低一半的處理時間,同時也可以降低分類錯誤的機率,提高其精準度。
| zh_TW |
dc.description.abstract | Remote sensing images offer us the information of the ground spectral data that could help us analyze and tell what the true land surface condition is. Nevertheless, because of the limited spectral information, multispectral remote-sensing images are difficult to be classified with high accuracy and efficiency especially in conventional classification methods.
We devised a method called affinity propagation which helps input measures of similarity between pairs of data points. Real-valued messages are exchanged between data points until a high-quality set of exemplars and corresponding clusters gradually emerges.
While analyzing, the efficiency is considered. Processing all data points simultaneously causes the slow processing speed because of the huge amount of data. Research here combines a hierarchical approach to find the best solution. In this experiment, we used hierarchical affinity propagation to cluster hyper spectral image from Airborne Imaging Spectrometer ROSIS. Experiment result showed that the processing time was cut into half, and the accuracy of the outcome was also enhanced.
| en_US |
DC.subject | 影像分類 | zh_TW |
DC.subject | 區域成長理論 | zh_TW |
DC.subject | 親和力傳播理論 | zh_TW |
DC.subject | Region Growing Method | en_US |
DC.subject | Affinity Propagation | en_US |
DC.subject | Classification | en_US |
DC.title | 應用階層式親和力傳播理論進行高光譜影像分類 | zh_TW |
dc.language.iso | zh-TW | zh-TW |
DC.title | Hierarchical Affinity Propagation Technique for Clustering in Hyperspectral Images | en_US |
DC.type | 博碩士論文 | zh_TW |
DC.type | thesis | en_US |
DC.publisher | National Central University | en_US |