利用耦合非負矩陣分解方法進行Hyper-SCAN影像合成;Hyper-SCAN image fusion using coupled non-negative matrix factorization

NCUIR > College of Earth Sciences > Graduate Institute of Space Science > Department of Earth Sciences > Item 987654321/84896

Please use this identifier to cite or link to this item: https://ir.lib.ncu.edu.tw/handle/987654321/84896

Title:	利用耦合非負矩陣分解方法進行Hyper-SCAN影像合成;Hyper-SCAN image fusion using coupled non-negative matrix factorization
Authors:	劉晏辰;Liu, Yen-Chen
Contributors:	太空科學與工程研究所
Keywords:	高光譜儀;影像融合
Date:	2021-01-28
Issue Date:	2021-03-18 16:48:21 (UTC+8)
Publisher:	國立中央大學
Abstract:	Hyper-Spectral Camera Analyzer (Hyper-SCAN) 為本實驗室自行組裝之高光譜儀，計畫將於2022年搭載於scion微衛星太空任務中，特色為體積小、重量輕以及成本較低，其連續光譜帶範圍在464nm~676nm，擁有高光譜分辨率(~1nm)，此儀器以推掃式(push-broom)掃描獲取高光譜影像，視角大約為5.6度，在高度500公里的掃描幅寬為48.9公里，中心波長為570nm。由於Hyper-SCAN仍在開發階段，獲取影像的過程中需要調整多種儀器，因此在本論文中提出一套取像實驗流程，並透過儀器的自動化控制減少整體實驗過程中的人工誤差與儀器操作時間。再者由於同時具有高空間解析度與高光譜解析度的光譜儀造價不斐且存在機構上的限制，因此文中使用耦合非負矩陣分解(CNMF)法融合高光譜影像資料與多光譜影像，同時亦對不同的初始端元(endmember)光譜作為演算法輸入所融合的結果進行比較。 ;Hyper-Spectral Camera Analyzer (Hyper-SCAN) is a self-assembled hyperspectrometer in this laboratory. It is planned to be installed in the scion microsatellite space mission in 2022. It is characterized by small size, light weight and low cost. Its continuous spectrum The band range is 464nm~676nm, with high spectral resolution (~1nm), this instrument uses push-broom scanning to obtain hyperspectral images, the viewing angle is about 5.6 degrees, and the scan width at a height of 500 kilometers is 48.9 Kilometers, the center wavelength is 570nm. Since Hyper-SCAN is still in the development stage, many instruments need to be adjusted in the process of acquiring images. Therefore, in this paper, a set of image acquisition experiment procedures is proposed, and the automatic control of the instruments reduces the manual error and instrument operation time in the overall experiment process. Furthermore, because spectrometers with both high spatial resolution and high spectral resolution are expensive and have institutional limitations, the coupled non-negative matrix factorization (CNMF) method is used to fuse hyperspectral image data and multispectral images. Different initial endmember spectra are used as input to the algorithm to compare the fusion results.
Appears in Collections:	[Graduate Institute of Space Science] Department of Earth Sciences

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	271	View/Open

社群 sharing

Loading...