摘要
逆光是造成图像质量变低的重要原因之一。针对逆光图像中亮度降低及细节损失等问题,论文在单尺度Retinex算法的基础上,采用视觉感知质量的图像评价方式对图像进行融合。在单尺度Retinex算法中,不同的高斯函数尺度参数会产生不同的平滑增强效果,使用一组尺度参数的高斯函数对图像进行增强,对获得的增强组图进行视觉感知质量评价,根据评价结果进行图像融合,使得图像的纹理细节部分更加突出,整体增强效果更好,平衡了全局与局部的增强性能。实验结果验证了论文所提出算法的有效性。
Backlighting is one of the important reasons for the low image quality.Aiming at the problems of brightness reduction and detail loss in backlight images,based on the single-scale Retinex algorithm,the image is fused by visual perception image quality evaluation.In the single-scale Retinex algorithm,Gaussian functions with different scales produce different smoothing enhancement effects.If the image is enhanced by a Gaussian function of a set of scale parameters,then the visual enhancement quality of the obtained enhanced group images is evaluated,and the image fusion is performed with the evaluation result.Image enhancement by this method can make the texture details of the image more prominent,and the overall enhancement effect is better,balancing the global and local enhancement performance.The experimental results verify the effectiveness of the proposed algorithm.
作者
李明刚
李海山
李恒
LI Minggang;LI Haishan;LI Heng(Wuhan Digital Engineering Institute,Wuhan 430074)
出处
《计算机与数字工程》
2020年第4期952-955,985,共5页
Computer & Digital Engineering
