摘要
针对带颜色恢复的多尺度Retinex算法在最后输出的图像上存有重叠的问题,提出了一种改进的子频带分解的Retinex算法,该算法不仅能增强亮点中的细节,也能增强在阴影中的细节。由于RGB这三种颜色之间有很强的关系,而HSV这三者之间没有很大的关联,能够较好的反映人对色彩的感知。实验结果表明,与基于RGB空间的多尺度Retinex算法相比,改进的基于HSV的算法更有效的增强图像在亮点和阴影中的细节,颜色更接近于原图。
For the muhiscale Retinex algorithm with color restoration on the final output images existing a problem of overlapping, we proposed an improved Retinex algorithm of sub-band decomposition. The algorithm can not only enhance the details in the highligh4 but also can improve the details in the shadows. Because the RGB has a strong relationship between these three kinds of color, however, there is no great connection in the HSV, the HSV are able to reflect the perception of color better. The experiments show that compared with the multiscale Retinex algorithm based on RGB space, the improved multiscale Retinex algorithm based on the HSV color space is more effective to enhance image details in the highlight and shadows and the color is more close to the original image.
出处
《电子设计工程》
2015年第7期148-150,共3页
Electronic Design Engineering
