土地覆蓋和土地利用的變遷偵測是遙測影像上一項重要的應用。影像變遷偵測有兩個重要的工作：(1)變遷區域之偵測(2)類別變化之偵測。過去在傳統上主要是以像元為基礎的方式來進行變遷區域偵測的工作，在偵測的結果上往往會出現許多如像元大小般的雜訊。因此，在本研究中提出以區塊為基礎的技巧，來進行變遷偵測的工作。本研究主要的方式是對多時期影像先進行合併的工作，再利用區塊增長法(region growing)對於合併後的影像進行區塊化的分割，接著再應用統計上的顯著水準測試來檢驗區塊有無變遷情形發生。最後，利用資料挖掘(data mining)之技術建立類別知識庫，來尋變遷區塊之類別變化情形。此方法實際運用在模擬影像及SPOT5影像上進行實驗的測試與評估，其結果顯示本研究所提出以區塊為基礎及資料挖掘技術的變遷偵測方式，可提供完整的變遷區塊及類別變遷的資料。 Land cover/land use change detection has been an important application of remotely sensed images. Basically, a practical change detection algorithm is expected to provide a complete information about the change that includes changed locations and categories. Unlike the pixel-based change detection methods, which normally generate pixel-sized noises, this study develops a segmentation-based change detection technique. The main idea of the study is to apply region growing segmentation to multi-dated composite images, and divide the images into many regions. Then a statistical significance test is used to find out the regions that have been changed or not. Thereafter, a knowledge-based system developed by the data mining technique is used to find out the category change for each changed region. An experiment is performed to detect the changed areas with simulated and SPOT5 images, and the result shows that the segmentation-based change detection method combined with data mining technique can provide region-liked changed map as well as the change information about the category.
