Underwater Image Enhancement Based on the Fusion of Matching Transformation and Linear Transformation
Aiming at better improvement of the color and contrast of underwater image,an underwater image enhancement method based on the fusion of matching transformation and linear transformation is proposed.The channel image with highest intensity is enhanced by the contrast limited adaptive histogram equalization,and based on the histogram of the equalized channel image,the other two channel images are enhanced by histogram matching,so as to obtain color corrected image.And in the HSV color space,the brightness component of the color corrected image is enhanced by adaptive piecewise linear transformation,so as to obtain contrast enhanced image.With the help of the contrast weight,saliency weight,brightness distribution weight,and saturation weight calculated separately with the two enhanced images,the two enhanced images are multi-scale pyramid weighted fused,obtaining the final enhanced image.Experimental results shown that the visual effect of enhanced images by proposed method outperforms the existing methods,and the corresponding average gradient,UCIQE and UIQM are higher than the existing methods by 30.0％,7.29％and 16.0％,respectively.
underwater image enhancementhistogram matchingadaptive piecewise linear transformationpyramid weighted fusion
万方数据
维普
