摘要
针对在海洋环境中采集到的水下图像由于光的散射与吸收作用导致的成像模糊以及退化的问题,提出了一种NSST与PCNN相结合的水下图像融合方法,在改变通道补偿方式的基础上分别进行Gamma校正得到对比度增强图像以及图像锐化处理得到细节信息增强后的图像,将两图像运用NSST变换分别得到低频子带系数和带通子带系数,使用基于区域能量自适应加权融合规则将低频方向子带系数融合;采用带通子带系数作为外部激励,刺激PCNN神经元点火,经脉冲发生区域选用像素较大的脉冲输出,使带通子带系数融合;对融合低频子带系数和带通子带系数分别做NSST逆变换,获得重构图像。通过实验结果以及客观评价可知,WNPF算法解决了水下图像色彩失真的问题,同时图像清晰度得到了提升。
For collection of images in the Marine environment caused by light scattering and absorption of image blur and degradation problems,an underwater image fusion method combining NSST and PCNN is proposed.On the basis of changing the channel compensation mode,the contrast enhanced image was obtained by Gamma correction and the image sharpening was obtained by detail enhanced image.The two images were decomposed into low-frequency subband coefficient and band-pass subband coefficient by NSST transform.The sub-band coefficients of low frequency direction are fused by using the adaptive weighted fusion rule based on regional energy.The band-pass subband coefficient was used as the external excitation to stimulate the ignition of PCNN neurons,and the pulse output with large pixel was selected through the pulse generation region to make the band-pass subband coefficient fusion.The reconstructed image was obtained by NSST inverse transformation of the fused low-frequency subband coefficient and band-pass subband coefficient.Through subjective observation and objective evaluation,WNPF algorithm solves the problem of underwater image color distortion and improves the image clarity.
作者
韩丽
王中训
陈玉杰
刘培学
王林霖
HAN Li;WANG Zhongxun;CHEN Yujie;LIU Peixue;WANG Linlin(College of Physics and Electronic Information,Yantai University,Yantai 264005,China;College of Intelligent Manufacturing,Qingdao Huanghai University,Qingdao 266555,China)
出处
《电子设计工程》
2024年第20期72-77,共6页
Electronic Design Engineering
基金
山东省重点研发计划(2017GGX201004,2019GGX105001)
西海岸新区高校校长基金(XZJJZY01)。
