理想直方图匹配与Canny边缘的水下图像增强  

Underwater Image Enhancement Based on Ideal Histogram Matching and Canny Edge

作　　者：尹小琴 罗启强 YIN Xiaoqin;LUO Qiqiang(School of Information and Artificial Intelligence,Nanchang Institute of Science and Technology,Nanchang Jiangxi 330108,China)

机构地区：[1]南昌工学院信息与人工智能学院,江西南昌330108

出　　处：《传感技术学报》2024年第12期2071-2077,共7页Chinese Journal of Sensors and Actuators

基　　金：江西省教育厅科学技术研究项目(GJJ2208506,GJJ2208520);南昌工学院科技计划博士专项项目(NGKJ-22-01);南昌工学院机器人与智能系统研究中心项目(NGYZY-20-005)。

摘　　要：针对水下图像增强后色彩不够自然、对比度低和过增强等问题,提出一种基于理想直方图匹配与Canny边缘的水下图像增强方法。将假设的均匀分布的直方图作为理想直方图,根据理想直方图对三通道图像分别进行直方图匹配,以校正图像的颜色和提升图像的对比度。进一步地,借助于Canny边缘算子,对三通道图像分别进行边缘定位,然后对其进行边缘融合增强,突出图像的边缘细节信息。实验结果显示,相对于部分现有方法,所提方法对水下图像增强后的视觉效果更佳、图像的信息熵、标准差和水下图像质量指标更高。因此,所提方法具有更优的水下图像增强效果。Targeting at the color unnaturalness,low contrast and over enhancement of the enhanced underwater image,an underwater im-age enhancement method based on ideal histogram matching and Canny edge is proposed.This method takes a hypothetical uniformly distributed histogram as the ideal histogram,based on which,three channel images are subjected to histogram matching,respectively,so as to correct the color of image and improve the contrast of image.Moreover,with the help of Canny edge operator,it performs edge lo-calization on the three channel images,respectively,and then performs edge fusion enhancement on them,so as to highlight the edge and detail of the image.The experimental results show that compared to some existing methods,the proposed method achieves better visual effect of image,higher values of information entropy,better standard deviation and high underwater image quality.Therefore,the pro-posed method has better enhancement effect for underwater image.

关 键 词：水下图像 图像增强 理想直方图 直方图匹配 Canny边缘增强 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]