水下无人系统学报2025，Vol.33Issue(1)：46-55,10.DOI:10.11993/j.issn.2096-3920.2024-0082
基于光照补偿与金字塔融合的水下图像增强方法
Underwater Image Enhancement Method Based on Illumination Compensation and Pyramid-Based Blending
摘要
Abstract
The existing enhancement methods based on deep learning and underwater imaging quality degradation models still suffer from poor robustness due to color deviation,scattering-induced blur,and uneven brightness in underwater optical imaging.In response to these issues,this article proposed a single image-based enhancement method for underwater imaging,which combined illumination compensation and pyramid-based detail blending.First,based on global illumination and color channel characteristics,illumination intensity at the pixel level was estimated and compensated,achieving intensity correction for each color channel.Next,the scatter components of the image were estimated using Gaussian blur,and a multi-scale Gaussian filter residual method was used for descattering.Finally,a multi-image pyramid-based detail blending technique was proposed,combining edge enhancement,adaptive Gamma correction,and brightness equalization.This method effectively preserved image details while addressing uneven brightness.Compared to existing techniques,this method featured improved adaptability and achieved better performance in underwater image quality measures(UIQMs)and underwater image color and quality evaluation(UCIQE)measures.关键词
水下图像增强/光照补偿/金字塔融合/亮度均衡/深度学习Key words
underwater image enhancement/illumination compensation/pyramid-based blending/brightness equalization/deep learning分类
武器工业引用本文复制引用
岳成海,徐会希,吕凤天,邵刚,朱宝彤,尹忠勋..基于光照补偿与金字塔融合的水下图像增强方法[J].水下无人系统学报,2025,33(1):46-55,10.基金项目
国家重点研发计划项目(2021YFC2800100) （2021YFC2800100）
广东省重点领域研发计划项目(2020B1111010004) （2020B1111010004）
辽宁省自然科学基金(2022-MS-035). （2022-MS-035）
