- 最后登录
- 2017-9-18
- 注册时间
- 2011-1-12
- 阅读权限
- 90
- 积分
- 12276
  
- 纳金币
- 5568
- 精华
- 0
|
1 Introduction
The camera response function determines the relationship between
the incident light on the camera sensor and the output pixel values
that are produced. For most consumer cameras, this function is proprietary
and needs to be estimated to create HDR images that accurately
represent the light distribution of the captured scene. Several
methods have been proposed in the literature to estimate this unknown
mapping using multiple exposures techniques. In this study,
we compare three of the most commonly used methods namely
Debevec and Malik’s [1997], Mitsunaga and Nayar’s [1999], and
Robertson et al.’s [2003] response curve estimation algorithms in
terms of how precisely they estimate an unknown camera response.
2 Comparison
We implemented all three algorithms in C++ using the exact procedures
outlined in the respective papers. We selected around 300
distinct sample positions from the uniform regions of the individual
exposures. In that we had 9 exposures, this amounted to a total of
2700 samples. We ensured that we had at least one sample for every
intensity level and the samples are not clumped together. A representative
set of samples is shown in Figure 1. As Mitsunaga and Nayar’s
algorithm does not have a weighting mechanism to underplay
the influence of under- and over-exposed samples, we discarded the
samples outside the range [5; 250] only for that algorithm.
The recovered response curves of a Canon EOS 550D using three
of the studied algorithms are shown on the right of Figure 1. As this
figure shows, the response curves are not only different from each
other but they also deviate significantly from the sRGB gamma. To
evaluate the fidelity of each algorithm, we devised the metric |
|
发表于 2011-12-29 08:59:17
QQ好友和群
腾讯微博
腾讯朋友
微信
转播
淘帖
收藏
支持
反对
发表于 2012-8-16 10:52:20