臺灣地暉星震觀測網（Taiwan Earthshine and AsteroSeismology Telescopes, EAST）將來可望可以提供大量的觀測數據。此研究希望利用這些數據，使用一個快速的分析方法，發現新的變星或掩星的存在。
變星是發光亮度會隨著時間而改變的恆星。其亮度變化被認為跟星震有關聯，所以亮度變化具有週期性。對現有觀測資料最多的盾牌座δ型變星來說，其週期由數分鐘到數天都有可能。
掩星也是亮度會隨著時間而改變的恆星，但其亮度變化的原因和變星不一樣，大致上可分為兩種：1.從地球的角度來看，掩星本身的行星運行軌道，剛好通過恆星和觀測者的中間，所以有時候該恆星的光會被行星擋住。2.太陽系內的隕石擋在恆星和觀測者的中間，擋住了恆星的光。
傳統上尋找變星或掩星的方法是測光分析。即一次針對一個目標星，分析其亮度隨時間是否有變化。此研究希望使用更快的方法，一次不只針對一個目標星，而是可以將觀測視野中所有的星星一起做分析。快速地判別出有可能是變星或掩星的目標之後，再對該目標仔細做測光分析。這樣即可大大減少做測光分析的次數，減少許多分析EAST大量數據的時間。
變星和掩星的共通點，就是亮度會隨著時間而改變，雖然變化的機制不一樣。所以若是計算恆星的亮度隨時間變化的標準差，就能快速地判別一般恆星和這兩種亮度會變化的恆星。標準差小的就是一般恆星，反之就可能是變星或掩星。
為了讓程式有辦法計算恆星亮度隨時間變化的標準差，有兩個重要的步驟。第一個是將影像對齊，讓同一顆星在影像上保持相同的座標。第二個主要的步驟是做空間上的移動平均。當平均範圍小於星點的大小，有讓影像變平滑的效果，但這裡的目的不是這樣。這裡是要讓平均範圍大於星點的大小，使得星點中央的一個數值可以用來代表整顆星的亮度，因為這個數值是星點總亮度加上一點點背景的亮度。這步驟就像是一個簡單的孔徑測光。
實際上用此方法分析數天的觀測數據，結果此分析方法可以判別出許多已知變星的存在。

Taiwan Earthshine and AsteroSeismology Telescopes (EAST) System is to study long-term variation of the earthshine and asteroseismology; Taiwan EAST System therefore will provide a huge amount of data in the future. In this thesis, we develop a method which could be applied to the data of EAST to survey unidentified variable stars and occultation events.
The brightness of a variable star changes with time; the variation is caused by stellar oscillation. For δ scuti variable stars, the period rang from several minute to several days.
The causes of occultation events are: (1) a planet passes through the line of sight between the host star and the observer, therefore caused a sudden drop of the host star’s brightness. (2) meteorites pass through the line of sight between the observer and a star, and blocking the star light from the observer.
Usually, we analyze variation stars or occultation events by applying aperture photometry. However, aperture photometry could only be applied to one target star each time, and is less effective to search unidentified variable stars or occultation events from mass EAST data. In this research, we hope to develop a more effective method to survey EAST data and identify possible variable star candidates. Of course, further analysis should be made after identifying variable star candidates.
The brightness of variable stars and occultation events vary with time, although the mechanism remains different. Therefore, brightness variation could be identified from calculating the standard variation of star’s brightness with respect to time. If the standard variation of a star is bigger than other stars, it could be a variable star.
There are two important steps to calculate the standard variation of star’s brightness. First, align images of an observation night. Second, calculate the 2-D spatial running mean from images. [If the area of running mean is smaller than a star in the image, we get a smoother image; however, this deviates from our research purpose. In our case, we choose the area of running mean bigger than a star. It turns out that the value at star’s center could be regard as the brightness of the whole star, for the reason that this center value is the summation of the star’s brightness and nearby background. In other words, it is like a simple aperture photometry.]
We apply this method to several days’ observation of EAST, and the results show that several known variable stars could be detected perfectly.

[1] 張靖歆（2006，清大碩士論文）
[2] 王立涵（2007，清大碩士論文）
[3] Numerical Recipes in C, 2nd, Cambridge Univ. Press
[4] M. Breger, R. Garrido, G. Handler, M.A. Wood, R.R. Shobbrook, K.M. Bischof, F. Rodler, R. O. Gray, A. Stankov, P. Martinez, D. O’Donoghue, R. Szabo’, W. Zima, A. B. Kaye, C. Barban, U. Heiter (2002, MNRAS) 29 frequencies for the δ Scuti variable BI CMi: the 1997-2000 multisite campaigns