高動態範圍之影像可使用不同的方法生成,最常見的一個方法是透過許多不同動態範圍的影像來重建相機響應曲線(CRF),將數值延伸成高動態範圍值域,再經過色調對應,將數值對應回顯示器所能顯示的動態範圍。
數位相機上的EV值取決於光圈、快門、iso的組合,可配合不同的場景需求,來設定不同的數值,以得到相同的場景輝度值,但是無法確定數位相機上的快門時間是否為它顯示般的準確,會導致場景輝度上的偏差,由於物體的色度是不變的,而輝度所使用單位都是基於cd⁄m^2 這樣的實際物理單位,透過校正輝度能夠更加精準的重現真實場景,也能在對應回數位影像顯示設備上更加精準。
本研究先將相機不同曝光設定產生的數值轉換成三色刺激值XYZ,再透過迴歸分析法來找出轉換後的數值與真實場景量測之輝度值的關係。迴歸分析法,能找出一個最能夠代表所有觀測資料的函數、判斷兩變數之間的關係,只要依變數與自變數之間存在著相當比率的關係,則稱為兩條向量曲線擁有線性關係,就能找出其中的轉換函式,並產生不同曝光設定的修正函式,以修正相機在拍攝上的不足、讓高動態範圍還原技術更加精準。
There are many ways to obtain a high dynamic range image. The most common algorithm is to rebuild the camera response function to high dynamic range radiance maps from a series of differently exposed images of the same scene. After merging process, one set of colors to another to approximate the appearance of high dynamic range images in a medium that has a more limited dynamic range by using image processing and computer graphics to map.
Digital cameras' exposure value is depended by aperture, Shutter and Iso setting. One can get the same scene radiance value from different setting, but can't make sure the influence of shutter setting in radiance value produce is correct. Through correct the brightness of objects can rebuild the scene radiance precisely.
Regression Analysis is a statistical process for estimating the relationships among variables. This research proposes an algorithm by regression analysis to recover high dynamic range radiance maps from a series of differently exposed images of the same scene.