Journal of Computer Applications ›› 2019, Vol. 39 ›› Issue (9): 2738-2743.DOI: 10.11772/j.issn.1001-9081.2019020353

• Virtual reality and multimedia computing • Previous Articles     Next Articles

Underwater image super-resolution reconstruction method based on deep learning

CHEN Longbiao, CHEN Yuzhang, WANG Xiaochen, ZOU Peng, HU Xuemin   

  1. School of Computer Science and Information Engineering, Hubei University, Wuhan Hubei 430062, China
  • Received:2019-03-05 Revised:2019-05-05 Online:2019-09-10 Published:2019-06-03
  • Supported by:

    This work is partially supported by Youth Science Foundation of National Natural Science Foundation of China (61806076), the Student's Platform for Innovation and Entrepreneurship Training Program of Hubei Province (201710512051, 201810512051).

基于深度学习的水下图像超分辨率重建方法

陈龙彪, 谌雨章, 王晓晨, 邹鹏, 胡学敏   

  1. 湖北大学 计算机与信息工程学院, 武汉 430062
  • 通讯作者: 谌雨章
  • 作者简介:陈龙彪(1997-),男,湖北咸宁人,主要研究方向:深度学习、图像处理;谌雨章(1984-),男,湖北武汉人,副教授,博士,主要研究方向:光电探测、图像处理;王晓晨(1998-),男,河南郑州人,主要研究方向:深度学习、软件工程;邹鹏(1997-),男,湖北鄂州人,主要研究方向:图像处理、深度学习;胡学敏(1985-),男,湖南岳阳人,副教授,博士,主要研究方向:计算机视觉、智能系统。
  • 基金资助:

    国家自然科学基金青年科学基金资助项目(61806076);湖北省大学生创新创业训练计划基金资助项目(201710512051,201810512051)。

Abstract:

Due to the characteristics of water itself and the absorption and scattering of light by suspended particles in the water, a series of problems, such as low Signal-to-Noise Ratio (SNR) and low resolution, exist in underwater images. Most of the traditional processing methods include image enhancement, restoration and reconstruction rely on degradation model and have ill-posed algorithm problem. In order to further improve the effects and efficiency of underwater image restoration algorithm, an improved image super-resolution reconstruction method based on deep convolutional neural network was proposed. An Improved Dense Block structure (IDB) was introduced into the network of the method, which can effectively solve the gradient disappearance problem of deep convolutional neural network and improve the training speed at the same time. The network was used to train the underwater images before and after the degradation by registration and obtained the mapping relation between the low-resolution image and the high-resolution image. The experimental results show that on a self-built underwater image training set, the underwater image reconstructed by the deep convolutional neural network with IDB has the Peak Signal-to-Noise Ratio (PSNR) and Structural SIMilarity (SSIM) improved by 0.38 dB and 0.013 respectively, compared with SRCNN (an image Super-Resolution method using Conventional Neural Network) and proposed method can effectively improve the reconstruction quality of underwater images.

Key words: Convolutional Neural Network (CNN), super-resolution reconstruction, Signal-to-Noise Ratio (SNR), underwater image processing, mapping

摘要:

由于水体本身的特性以及水中悬浮颗粒对光的吸收和散射作用,水下图像普遍存在信噪比(SNR)低、分辨率低等一系列问题,但大部分方法传统处理方法包含图像增强、复原及重建,都依赖退化模型,并存在算法病态性问题。为进一步提高水下图像恢复算法的效果和效率,提出了一种改进的基于深度卷积神经网络的图像超分辨率重建方法。该方法网络中引入了改良的密集块结构(IDB),能在有效解决深度卷积神经网络梯度弥散问题的同时提高训练速度。该网络对经过配准的退化前后的水下图像进行训练,得到水下低分辨率图像和高分辨率图像之间的一个映射关系。实验结果表明,在基于自建的水下图像作为训练集上,较卷积神经网络的单帧图像超分辨率重建算法(SRCNN),使用引入了改良的密集块结构(IDB)的深度卷积神经网络对水下图像进行重建,重建图像的峰值信噪比(PSNR)提升达到0.38 dB,结构相似度(SSIM)提升达到0.013,能有效地提高水下图像的重建质量。

关键词: 卷积神经网络, 超分辨率重建, 信噪比, 水下图像处理, 映射

CLC Number: