遥测应用於土地利用分类.pdfVIP

遙測應用於土地利用分類- 以美濃中壇為例 Remote-Sensing Application on Land Cover Classification ―A Case Study at Meinung Zhang Tan Area 1 2 3 4 5 6 陳鈞華 、童伯倫 、方彥凱 、譚智宏張煜權 、甘俊二 1.屏東科技大學副教授 2.屏東科技大博士班 3.屏東科技大學碩士 4.農工中心副研究員兼組長 5 立德管理學院助理教授 5 七星農田水 利基金會董事長 摘 要 本研究選定美濃中壇工作站為研究區域，利用遙測技術藉由近紅外光與 短紅外光在 SPOT 4衛星所擁有的特性，探討兩波段應用在常態化植生指數 標準差 (NDVI)在土地利用分類的準確度增進效果。由兩張不同時期影像的水 稻分類準確度評估中短波 NDVI的水稻使用者準確度達 85.71 % 短波 NDVI的水稻使用者準確度達 87.10 % ，短波紅外光 波段應用於常態化植生指數標準差較近 紅外光為佳。討論選用的兩個 NDVI 波段與分類間之相關性，結果顯示短波 NDVI 於各分類類別的相關性較不顯 著，故有助於分類時準確度的提升。 關鍵詞：常態化值生指數標準差，波譜值，相關性 Abstract The study area in this research is in Meinung Zhang Tan work station. Utilizing remote sensing technique that according the near infrared and the short-wave infrared bands of SPOT 4 satellite. Those two bands were applied to the Normalized Difference Vegetation Index (NDVI) and compared their contribution in the classification accuracy in two rice growing periods. Results indicated that the classification accuracy of the short-wave infrared NDVI is better than that of near infrared. The User’s Accuracy of the paddy classification is 85.71 % using the short-wave NDVI in the image of Feb. 27, 2003. The User’s Accuracy of the paddy is 87.10 % using the short-wave NDVI in the image of Jul. 19, 2003. The result shows that short-wave NDVI is less dependent than other bands in the classification. In other words, the short-wave NDVI is a better band to be used in classification. Keywords: Normalized Difference Vegetation Index ，spectrum value ， Correlation -80-