光学学报, 2017, 37 (9): 0910002, 网络出版: 2018-09-07   

基于Tetrolet变换的彩色水下图像清晰化算法 下载: 963次

A Color Underwater Image Clearness Algorithm Based on Tetrolet Transform
作者单位
兰州交通大学电子与信息工程学院, 甘肃 兰州 730070
摘要
为了对水下图像传感器获取的彩色图像进行清晰化处理,提出一种lαβ色彩空间的图像清晰化算法。将捕获到的彩色水下图像进行暗原色初步复原后,映射RGB空间至lαβ三通道进行清晰化处理。采用Tetrolet变换方法处理亮度通道l,对包含大部分的轮廓、边缘等线性细节的高频分量采用Bilateral滤波器处理,对包含大部分能量的低频分量进行非局部均匀滤波处理,然后将处理后的高、低频分量进行反向Tetrolet变换得到复原的亮度通道图;对αβ通道的色彩偏差进行空间均值颜色校正,得到复原的αβ通道图。将处理后的非彩色通道图和彩色通道图反向变换至RGB通道,更新透射率,得到清晰化的彩色水下图像。实验结果表明,该算法对彩色水下图像的复原效果较好,在图像色彩的提升和边缘细节的描述方面效果显著。
Abstract
For the clearness of colored images captured by underwater image sensors, a lαβ color space image clearness algorithm is proposed. On the basis of former dark channel prior algorithm treatment on captured color images, the space mapping is used for further clearness processing, and changing RGB images to lαβ color space images. For the brightness channel l, the Tetrolet transform method is adopted. The bilateral filter is used to filter underwater image noise of high-frequency components, which contains most rough, edges, and other linear details. The nonlocal uniform filter is used to recover the clear low-frequency components, which contains the image’s most energy. Processed high-frequency and low-frequency images are inversely converted by Tetrolet transform to get recovered brightness channel image. The spatial mean color correction method is adopted to get recovered α and β channel images, and the processed brightness channel image and color channel images are inversely transformed to RGB space. Updating the transmissivity by input image to get clear color underwater image. Experimental results show that the proposed algorithm works well on underwater image restoration, and it is effective on the image color promotion and edge description.

沈瑜, 党建武, 王阳萍, 王博伟. 基于Tetrolet变换的彩色水下图像清晰化算法[J]. 光学学报, 2017, 37(9): 0910002. Yu Shen, Jianwu Dang, Yangping Wang, Bowei Wang. A Color Underwater Image Clearness Algorithm Based on Tetrolet Transform[J]. Acta Optica Sinica, 2017, 37(9): 0910002.

本文已被 3 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!