基于图像熵最小化的渐晕校正方法

Vignetting correction method based on minimizing image entropy

渐晕效应是指在光学成像系统中光线到达成像面上的光强度随着光线和光轴间的距离增加而逐渐减弱的现象。为了校正该现象所造成的图像亮度失真,针对传统函数逼近法存在的校正不准确、鲁棒性差的问题,提出一种基于图像熵最小化的渐晕校正方法。首先,根据渐晕分布特性构造出合适的亮度映射函数,使用该函数对渐晕图像进行映射后计算其图像熵;然后,将该图像熵作为渐晕校正的衡量标准,通过最小化图像熵来拟合校正函数;最后,完成渐晕校正。实验表明:本文方法所拟合的校正函数更加准确,校正结果在结构相似性(SSIM)和均方根误差(RMSE)评价指标上优于所比较的方法,此外,本文方法能够有效校正不同场景、不同程度的渐晕效应,具有较强的鲁棒性。

The vignetting effect refers to the phenomenon that the intensity of the light reaching the imaging surface in the optical imaging system gradually weakens as the distance between the light and the optical axis increases.In order to correct the image brightness distortion caused by this phenomenon, a vignetting correction method based on image entropy minimization is proposed to solve the problem and poor robustness in the traditional function approximation method.Firstly, a suitable brightness mapping function according to the vignetting distribution characteristics is constructed.Then, this function is used to map the vignetting image and calculate its image entropy, which is served, as a measure of vignetting correction.The correction function is fitted based on the proposed entropy, by which vignetting correction is performed.Experiments show that the correction function fitted by the proposed method is more accurate, and the correction result is better than the existing methods in terms of structural similarity(SSIM)and root mean square error(RMSE)indicators.The proposed method can effectively correct vignetting effect, it has strong robustness.