Efficient Sandstorm Image Enhancement Using the Normalized Eigenvalue and Adaptive Dark Channel Prior

Description

The color casted image has imbalanced color channel. In case of underwater image, the greenish or bluish color appears due to attenuated color channel by scattering of lightness. The distribution of color channel of underwater image has abundant blue channel and sparse component of red channel. Meanwhile the sandstorm image has reddish or yellowish color cast due to scattering of colorfulness sand particles; the red channel is abundant and blue channel is sparse. Because the color casted image has imbalanced color channel, to enhance the casted image, image adaptive color balancing procedure is needed. The eigenvalue of image shows the feature of image. In case of red channel, the maximum eigenvalue of red channel is bigger than that of blue and green channels. Therefore because the eigenvalue of image is able to reflect the image’s feature as with contrast, this book shows the color balancing method using the image’s eigenvalue. The eigenvalue of image has different value according to condition of color channels. As mentioned above, if the image is attenuated more, then the maximum eigenvalue is lower and vice versa.