High Dynamic Range Image Acquisition from Multiple Low Dynamic Range Images Based on Estimation of Scene Dynamic Range Kee-Hyon Park, Dae-Geun Park and Yeong-Ho Ha  School of Electrical Engineering and Computer Science, Kyungpook National University, 1370 Sankyuk-dong, Buk-gu, Daegu 702-701, Korea E-mail: yha@ee.knu.ac.kr Abstract. To acquire a high dynamic range (HDR) image of a scene, several low dynamic range (LDR) images acquired from a digital camera with different exposure times are generally fused into one HDR image to cover the entire dynamic range of the scene. However, when capturing a scene, the scene dynamic range (SDR) is unknown. Consequently, the exposure times for the LDR images need to be as varied as possible to cover the unknown SDR. This paper proposes a method to estimate the SDR using two LDR im- ages. Using the SDR information, SDR-adaptive exposure times can then be selected to achieve the optimal HDR image. The SDR is defined as two exposure times when captured LDR images are mar- ginally clipped to black and white, indicating the lower and upper limits of the SDR, respectively. To identify these times, two LDR images, an overexposed and an underexposed image, are cap- tured. Using the opto-electronic conversion function of the camera used, the minimum gray level in the overexposed image is then used to estimate the exposure time to make the minimum gray level of the image just black, while the maximum gray level in the under- exposed image is used to estimate the exposure time to make the maximum gray level of the image just white. By evaluating the ac- quired HDR image error according to the exposure times of fused LDR images for various scenes, SDR-adaptive exposure times to acquire an optimal HDR image with the minimal error are selected. Experiments confirm that the quality of an HDR image based on fusing LDR images with the proposed SDR-adaptive exposure times is similar to that of an HDR image based on fusing LDR images with conventionally chosen exposure times, even though the number of LDR images used to acquire the HDR image with the proposed method is much smaller than that used by the conventional method. © 2009 Society for Imaging Science and Technology. DOI: 10.2352/J.ImagingSci.Technol.2009.53.2.020505 INTRODUCTION While improvements are consistently being made to the res- olution, image quality, design, and convenience of digital cameras, the dynamic range of the image sensors in digital cameras remains limited. Thus, when using a digital camera to take a picture of a scene that includes both bright and dark regions, the bright regions are often converted to white, while the dark regions are sometimes converted to black, as image sensors with a low dynamic range (LDR) cannot si- multaneously sense bright and dark information in an image with a high dynamic range. Thus, to overcome these limita- tions of the image sensors in digital cameras, research on the acquisition and reproduction of high dynamic range (HDR) images has attracted recent attention. The dynamic range that can be acquired using most digital cameras is limited, yet an HDR image can almost express the dynamic range of the real world, which is about 10 8 , whereas the dynamic range that the human eye can accommodate in a single view is about 10 5 , and the dynamic range of the image sensors in most digital cameras is about 10 3 , representing only 1 / 100 000 of the real world. The overall brightness information of the real world with a high dynamic range cannot be expressed using the normal 24-bit RGB image format with 8 bits per channel. Therefore, a new image format suitable for HDR images is required, such as 32-bit RGBE image format with additional 8-bit exponent information. 1 In addition, the reproduction of HDR images on a normal display with a low dynamic range is one of the most important challenges for high dynamic range imaging. As the dynamic range of a normal display is about 10 3 and such displays can only show image files with a 24-bit RGB format, HDR images need to be converted to a 24-bit RGB format using tone mapping or tone reproduction. 2,3 In general, there are two ways to acquire an HDR image: using a special HDR camera system that can accommodate the entire scene dynamic range or fusing several LDR images with multiple exposure times taken using a regular digital camera. Mann, Debevec et al., Robertson et al., and Mitsunaga et al. 4–7 already proposed various HDR image acquisition methods using LDR images captured using a regular digital camera, where a number of LDR images with multiple exposure times are fused to cover a broad range of both bright and dark regions, as the dynamic range of a real scene cannot be expressed by a single image taken using a regular LDR digital camera. An LDR image captured with a short exposure time can express bright areas, whereas a long exposure time can express dark areas. Therefore, capturing various images with different exposure times to cover the entire scene dynamic range (SDR) is important in all these methods. However, when a photographer captures a series of LDR images to acquire an HDR image of a scene, if the SDR is  IS&T Member. Received Oct. 6, 2008; accepted for publication Dec. 23, 2008; published online Mar. 16, 2009. 1062-3701/2009/532/020505/12/$20.00. Journal of Imaging Science and Technology® 53(2): 020505–020505-12, 2009. © Society for Imaging Science and Technology 2009 J. Imaging Sci. Technol. Mar.-Apr. 2009 020505-1