基于小波清晰度计算的水下图像融合增强研究

Underwater multi-focus image enhancement based on wavelet

近年来,机器视觉技术在水下环境探测中得到了广泛的运用,然而由于受到水下环境和光学系统景深的限制,无法获取清晰的全景图像。鉴于此,提出了一种基于小波清晰度计算的水下图像融合增强方法。首先,对源图像进行小波变换,并利用小波系数获取图像的清晰度。其次,利用获取的清晰度对图像进行区域划分,得到图像的聚焦区域。再利用不同的融合规则分别对聚焦区域和非聚焦区域进行融合。最后,将聚焦区域和非聚焦区域进行合成,得到最终的融合图像。实验证明,所使用的融合方法对于实验环境较差的水下图像效果有这较大的提升,更加便于水下环境的检测与分析。

In recent years,the machine vision based object detection methods have been widely applied in the underwater environments.However,due to the serious underwater environments and the limit visual depth of the underwater optical system,the quality of underwater images is poor.Aiming to slove this problem and making the underwater image clear,this paper proposes a fusion method based on the wavelet.First,we transform the source image into wavelet domain and get the clarity image.Then,the clarity image is segmented and the clarity for each area is calculated to select the focus area.Then the focus area and non-focus area are fused respectively in different method.Finally,the focus area and non-focus area are fused,producing the enhanced image.The experiment result shows that the fusion method we proposed in this paper has the ability to enhance the multi-focus underwater image,contributing to the underwater object detection tasks.