基于光流导向特征的美术绘画图像色彩校正方法

Color correction method of art painting image based on optical flow oriented feature

美术绘画图像拍摄时受拍照角度、相机曝光参数等原因的影响,导致图像色调和明暗具有显著差异,因此提出基于光流导向特征的美术绘画图像色彩校正方法。使用光流导向特征计算图像像素值,并迭代优化更新光流域。建立最小化变量能量总值目标函数得到光流信息,有效区分图像色彩失真区域。应用retinex原理提取高光区域,按照色度距离和空间距离设定系数权重,最后使用分层色彩校正获得高质量的色彩校正图像。实验结果表明,所提方法对不均匀多光源色偏图像具有良好校正效果,能够在美术绘画图像数字化资源的有效应用中发挥重要作用。

Due to the influence of camera angle,camera exposure parameters and other factors,there are significant differences in image hue and light and shade.Therefore,a color correction method based on optical flow oriented feature is proposed.The image pixel value is calculated by using optical flow guiding feature,and the optical watershed is updated iteratively.The objective function of minimizing the total energy of variables is established to obtain the optical flow information and effectively distinguish the color distortion region of the image.The Retinex principle is applied to extract the highlight area,and the coefficient weight is set according to the chromaticity distance and spatial distance.Finally,the layered color correction is used to obtain the high quality color correction image.The experiment results show that the proposed method has a good correction effect on uneven multi-light source color deviation image,and can play an important role in the effective application of digital resources of art painting image.