台灣人工林面積佔全省森林面積之20%，如何有效地監測這些人工林，爲今後森林經營者應正視的問題。本文旨在探討應用SPOT衛星資料監測人工林之可行性，並以林試所六龜試驗林之台灣杉人工林爲例，供未來全省人工林監測之參考。研究重點包括林地分級在選取均質樣區之應用、樣區光譜資料之分析、林地分級與影像分類之整合，以及提出具林地分級觀念之分層分類方法。研究結果指出，應用林地分級技術於遙測影像分類上，除可協助選取均質的樣區之外，並可提高分類準確度，對提昇應用SPOT資料監測人工林之可行性有明顯的幫助。 Plantations occupy 20% of forest areas in Taiwan. To monitor them well has become an important issue for forest managers. This study investigated the feasibility of applying SPOT data to monitor plantations and took Taiwania plantations of the Liukuei experimental forest as an example reference for monitoring the island's plantations. The research content includes the application of forest land class-ification to selecting homogeneous training plots, the analysis of training plots, the integration of forest land classification and image classification, and a proposed stratified classification approach incorporating forest land classification.The research results indicate that forest land classification introduced into image classification can improve not only the selection of homogeneous plots but also the classification accuracy. This result will improve the feasibility of applying SPOT data to monitor plantations.