Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/442141
Title: | Unmixing and Segmentation of Hyperspectral Images Using Unsupervised Nonlinear and Silhouette Technique |
Researcher: | Kriti |
Guide(s): | Garg, Urvashi |
Keywords: | Computer Science Computer Science Software Engineering Engineering and Technology |
University: | Chandigarh University |
Completed Date: | 2022 |
Abstract: | The arena of HSI processing is an application field for several methods. Amongst them, newlinehyperspectral unmixing offers a form of somatic image archetypal with simple elucidation newlineallowing subpixel resolution outcomes. It amounts to the identification of a position of spectral newlinesignatures that are pure and therefore called endmembers and their matching fractional, newlinedraftrulesabundances for every pixel in HSI. The proposed non-negative minimum volume newlinefactorization (NMVF) method yields better performance compared to the pure pixel-based newlinealgorithm. However, the existing algorithms focus on the three major steps for spectral unmixing newlinechain: 1) to estimate the endmembers count in a scene; 2) identify spectral signatures of newlineendmembers; 3) estimate fractional abundance for every endmember in every pixel of a scene. newlineMoreover, all the stages are performed by only a few algorithms in the process of hyperspectral newlineunmixing. The proposed method is different from other conventional methods as it begins with newlinethe overestimation of the count of endmembers wherein removing the endmembers that are newlineredundant by the means of collaborative regularization. newlineThe thesis also provides a hierarchically organized structure of algorithms existing in the literature newlinefor spectral unmixing. In HSI analysis, spectral unmixing is a tool. For this analysis, a requisite is newlinethe endmember s determination. In this thesis, the existing approaches support the endmember newlineorientation from the image statistics. The endmembers are anticipated to have specific physical newlinesignificance, probably in the case of methodologies that achieve an assortment from the newlineillumination pixel spectra. Though the methods typically yield convex polytopes to cover newlinealtogether the points in image statistics, therefore, the candidate customary of endmembers does newlinenot adequate in the official explanation of endmembers. Moreover, to increase the performance of newlineHU, spatial information incorporation has achieved great success. The thesis gives a brief of the newlineexis |
Pagination: | |
URI: | http://hdl.handle.net/10603/442141 |
Appears in Departments: | Department of Computer Science Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01_title.pdf | Attached File | 217.97 kB | Adobe PDF | View/Open |
02_prelim pages.pdf | 427.12 kB | Adobe PDF | View/Open | |
03_content.pdf | 248.19 kB | Adobe PDF | View/Open | |
04_abstract.pdf | 243.8 kB | Adobe PDF | View/Open | |
05_chapter 1.pdf | 1.45 MB | Adobe PDF | View/Open | |
06_chapter 2.pdf | 628.18 kB | Adobe PDF | View/Open | |
07_chapter 3.pdf | 566.5 kB | Adobe PDF | View/Open | |
08_chapter 4.pdf | 1.34 MB | Adobe PDF | View/Open | |
09_chapter 5.pdf | 350.62 kB | Adobe PDF | View/Open | |
10_annexures.pdf | 520.47 kB | Adobe PDF | View/Open | |
80_recommendation.pdf | 357.46 kB | Adobe PDF | View/Open |
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: