摘要
详细介绍了基于同一场景不同曝光量图像序列的高动态范围图像合成过程;针对现有拟合的相机响应曲线随采样点的变化曲线变化大、不稳定等特点,提出一种新的相机响应曲线算法。该算法先根据图像序列成像特点对图像序列去噪;采用最小二乘法拟合出图像序列像素值关系曲线;利用该曲线再一次采用最小二乘法拟合得到相机响应曲线。实验结果表明,该技术拟合出的相机响应曲线比较平滑,能合成出高质量的高动态范围图像。
A method of constructing high dynamic range images based on photographs of the same scene with different exposure is introduced. Aim at the recovered response curve is unstabitily and unsmooth, a new method of recovering the response curve is presented： firstly, denoise the sequential images; Then polynomial fit the curve between photographs using the least square method; finally, using this curve to recover the response curve by using the least square method again. The experimental results show that the response curve is smooth, and the constructed high dynamic range image is satisfactory.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第3期1032-1036,共5页
Computer Engineering and Design
基金
贵州省省长基金项目(黔省专合字(2007)14号)
关键词
高动态范围图像
相机响应曲线
采样点
最小二乘法
曝光度
high dynamic range image
response curve
selected pixels
the least square method
exposure
