Application of Support Vector Machines to airborne hyperspectral image classification
碩士 === 國立中興大學 === 土木工程學系所 === 96 === Airborne hyper-spectral remote sensing relative to the traditional techniques of remote sensing can acquire real-time information of the surface of the earth with less influence from cloud obstruction due to its lower flying height, and has been broadly applied i...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2008
|
Online Access: | http://ndltd.ncl.edu.tw/handle/61820302757536982069 |
id |
ndltd-TW-096NCHU5015102 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-096NCHU50151022016-05-09T04:13:51Z http://ndltd.ncl.edu.tw/handle/61820302757536982069 Application of Support Vector Machines to airborne hyperspectral image classification 利用支持向量機於機載高光譜感測影像之分類 Kai-Shiang Huang 黃凱翔 碩士 國立中興大學 土木工程學系所 96 Airborne hyper-spectral remote sensing relative to the traditional techniques of remote sensing can acquire real-time information of the surface of the earth with less influence from cloud obstruction due to its lower flying height, and has been broadly applied in target detection and classification due to its better spectral resolution. In this research, an airborne hyper-spectral image, which has 218 bands within a range of spectral resolution from 427.2nm to 945.7nm, is used to classify the vegetation of Mount Jou-Jou. Firstly, Normalized Difference Vegetation Index (NDVI) is used to differentiate the vegetation areas from non-vegetation areas. However, redundant bands could not significantly increase the accuracy of vegetation classification, but increase the computation cost of pattern recognition. Thus, the dimension of the hyper-spectral image is reduced by using Principle Component Analysis (PCA) to extract the useful information for vegetation classification. Finally, Minimum Distance to Mean Classifier (MDC), Guassian Maximum Likelihood Classifier (GML) and Support Vector Machines (SVMs) are employed to classify the vegetation based on the extracted useful information. In order to illustrate the classification accuracy of the above procedure, a hyper-spectral image of Purdue’s Indian Pines test site with its ground truth data is tested. Compared to 74.5% resulted from GML and 69.4% resulted from MDC, the best classification accuracy is 80% resulted from SVM. 楊明德 2008 學位論文 ; thesis 62 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立中興大學 === 土木工程學系所 === 96 === Airborne hyper-spectral remote sensing relative to the traditional techniques of remote sensing can acquire real-time information of the surface of the earth with less influence from cloud obstruction due to its lower flying height, and has been broadly applied in target detection and classification due to its better spectral resolution. In this research, an airborne hyper-spectral image, which has 218 bands within a range of spectral resolution from 427.2nm to 945.7nm, is used to classify the vegetation of Mount Jou-Jou. Firstly, Normalized Difference Vegetation Index (NDVI) is used to differentiate the vegetation areas from non-vegetation areas. However, redundant bands could not significantly increase the accuracy of vegetation classification, but increase the computation cost of pattern recognition. Thus, the dimension of the hyper-spectral image is reduced by using Principle Component Analysis (PCA) to extract the useful information for vegetation classification. Finally, Minimum Distance to Mean Classifier (MDC), Guassian Maximum Likelihood Classifier (GML) and Support Vector Machines (SVMs) are employed to classify the vegetation based on the extracted useful information. In order to illustrate the classification accuracy of the above procedure, a hyper-spectral image of Purdue’s Indian Pines test site with its ground truth data is tested. Compared to 74.5% resulted from GML and 69.4% resulted from MDC, the best classification accuracy is 80% resulted from SVM.
|
author2 |
楊明德 |
author_facet |
楊明德 Kai-Shiang Huang 黃凱翔 |
author |
Kai-Shiang Huang 黃凱翔 |
spellingShingle |
Kai-Shiang Huang 黃凱翔 Application of Support Vector Machines to airborne hyperspectral image classification |
author_sort |
Kai-Shiang Huang |
title |
Application of Support Vector Machines to airborne hyperspectral image classification |
title_short |
Application of Support Vector Machines to airborne hyperspectral image classification |
title_full |
Application of Support Vector Machines to airborne hyperspectral image classification |
title_fullStr |
Application of Support Vector Machines to airborne hyperspectral image classification |
title_full_unstemmed |
Application of Support Vector Machines to airborne hyperspectral image classification |
title_sort |
application of support vector machines to airborne hyperspectral image classification |
publishDate |
2008 |
url |
http://ndltd.ncl.edu.tw/handle/61820302757536982069 |
work_keys_str_mv |
AT kaishianghuang applicationofsupportvectormachinestoairbornehyperspectralimageclassification AT huángkǎixiáng applicationofsupportvectormachinestoairbornehyperspectralimageclassification AT kaishianghuang lìyòngzhīchíxiàngliàngjīyújīzàigāoguāngpǔgǎncèyǐngxiàngzhīfēnlèi AT huángkǎixiáng lìyòngzhīchíxiàngliàngjīyújīzàigāoguāngpǔgǎncèyǐngxiàngzhīfēnlèi |
_version_ |
1718262552672075776 |