查詢結果分析
相關文獻
- 空載DAIS高光譜資料於濱海地區土地利用分類上之應用
- 影像參數加權統計法PN指標於土地利用型高光譜特徵選粹上之研究
- 森林植被與地形因子對TM光譜資訊影響之研究
- State-of-the-art Technologies of Forest Inventory and Monitoring in Taiwan
- 遙測推估南臺灣不同土地利用溫度、植生指標之研究
- 動態學習法則類神經網路法在闊葉林區土地利用型分類上應用之研究
- 遙測資訊在生態環境變遷監測上應用之研究--以臺南地區為例
- 都市邊緣地區土地利用變遷之逕流特性研究--以景美溪集水區為例
- 高光譜主軸轉換影像辨識土地利用型最適主軸數決定方法之研究
- 應用主軸轉換法辨認空載高光譜影像土地利用型特性之研究
頁籤選單縮合
題名 | 空載DAIS高光譜資料於濱海地區土地利用分類上之應用=Application of Airborne DAIS Hyperspectral Data on Land Uses Classification of Marine Area |
---|---|
作者姓名(中文) | 林金樹; | 書刊名 | 國立臺灣大學農學院實驗林研究報告 |
卷期 | 15:1=231 2001.03[民90.03] |
頁次 | 頁15-24 |
分類號 | 554.5 |
關鍵詞 | 高光譜資料; 土地利用; 地類; 遙測; 濱海地區; Hyperspectral data; Land use; Land cover; Remote sensing; Marine area; |
語文 | 中文(Chinese) |
中文摘要 | 本文利用1997年行政院農業委員會之高光譜遙測計畫所取得的DAIS高光譜資料,探討空載高光譜資料檢測濱海地區土地利用型之效能。DAIS高光譜資料係由林務局農航所PA-31型飛機,搭載美國GER公司之DAIS-3715輻射光譜成像儀,於海拔1,828.8m (6,000ft)高空向地面偵測地物反射的光譜輻射能量所得到的;其可見光區和近紅外光區波段之光譜波長距離約20nm,輻射解析力為16位元,空間解析力為6m一個像元,遠優於Landsat(8-bit, 30-meter)及SPOT(8-bit, 20-meter)衛星多光譜資料。根據DAIS可見光區和近紅外光區共24個波段資料之分類結果,其訓練樣區之全區分類準確度(OA)和Kappa同意度係數(?)均高於98%,而評估樣區之OA和?也可達到89%和86%,顯示DAIS高光譜資料可以準確的檢測出45個細類的土地利用型,而且可以精確地分辨出有機物與無機物污染的水質;此等結果顯示,具高輻射解析力與高空間解析力之高光譜遙測資料,極適合應用於監測環境與區域土地利用之發展。 |
英文摘要 | The Council of Agriculture of Republic of China and GER Company in USA had jointly conducted an airborne hyperspectral remote sensing program referred as HYPER program that carried a DAIS-3715 spectroradiometer to image several typical ecozones of Taiwan. This study used one product of HYPER program to examine if the hyperspectral data was suitable for monitoring land uses in marine area. The wavelength interval for visible and infrared bands was about 20 nanometer. Objects reflected radiance was coded as 16 bits format. The flight altitude of HYPER program was 1,828.8m (6,000ft) feet and the IFOV of DAIS-3715 spectroradiometer was 3.3 miliradian, and hence produced a 6-mter spatial resolution data. All of the spectral, spatial, and radiometric resolutions of DAIS data were much better than Landsat and SPOT multispectral data. Accroding to the classification accuracy assessed from the training and assessing data sets, using 24-bands DAIS data was very powerful for detecting the land uses information. Both overall accuracy and kappa coefficient of agreement for the training data set was more than 98%, and those accuracy index values for the assessing data set were more than 89% and 86%, respectively. Organic and inorganic polluted water could also be identified exactly in this study. All of the results thus supported that a higher spatial and radiometric resolution of hyperspectral data was very effective to monitor the environmental changes and development of local land uses. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。