台灣位處於亞熱帶地區,是個海島型的國家,境內有很多超過一千公尺以上的高山,加上夏天對流旺盛以及颱風的影響,所以在台灣地區天空中常常會有雲霧的生成。因此對於光學衛星拍攝地表的影像會產生很大的影響。因為光學衛星的光波長無法穿透雲霧,導致地表資訊被雲霧所遮蔽,因此目的要將光學衛星影像上的雲霧濾除,以產生一張乾淨無雲的影像。本研究選取福爾摩沙二號的衛星影像作為研究影像,選取位於台灣東北角的地區包含雲的影像。首先先利用色彩轉換將紅、綠、藍色彩空間轉換成色調、飽和及亮度的色彩空間,用將雲在影像中加以凸顯。接著在影像中選取訓練樣本,利用監督式分類法來分類雲及陰影,在此運用最小歐基里德及城市區塊距離來做分類,分類完成後再利用型態學的方法來修復分類結果的破損部分,以及濾除雜訊。最後用分類結果來計算雲與陰影的相關係數,來找對應雲的陰影所在,並根據影像拍攝時的太陽仰角及方位角,還有兩者相對位置,估算雲的高度。此外,我們也運用白化-反白化矩陣去校正兩張不同時期的衛星影像,在雲以及陰影濾除後做影像鑲嵌,產生一張無雲及陰影的乾淨影像。 Taiwan is an island located at monsoon climate section in the Pacific Ocean. There are often clouds in the sky which poses restrictions for the development of optical remote sensing. Cloudless images over the island will take months to retrieve. For the sensors of visible and infrared light on earth resource satellites, the interference of cloud becomes a serious issue to deal with. In this case, we have to locate the cloud from the satellite image. We transfer image from (Red, Green, Blue) RGB system to the (Hue, Saturate, Intensity) HSI system to locate clouds. Then the clouds and shadow can be classified on the satellite image by supervised pure-pixel classification. We choose the city block, and Euclidean distance. We apply to classify the clouds and shadow. After this, the morphology to renovate the classify result and reduce noise. Finally we apply Image Color Correction for images collected of different dates and replace the cloud with mosaic process, and produce the cloudless images in an automatic sense. We can also calculate the correlation between cloud and its shadow to estimate the cloud height