摘要
针对水下图像退化现象严重、有效信息提取困难等问题,提出了水下退化图像处理方法.该方法通过分析水下图像退化过程,提出了基于大气湍流模型获取水下图像退化函数的方法,并利用频域滤波完成了退化图像的复原工作;进而将人工鱼群优化算法与图像二维Abutaleb熵信息相结合,利用一种二维最大熵阈值分割算法进行图像有意义区域分割.由于人工鱼群算法不需要了解问题的特殊信息,只进行问题优劣比较,使得该算法自适应性和收敛速度得到大幅提升.水池实验结果表明:该方法明显改善水下退化图像模糊度高、对比度低的问题,具有较优的分割效果,处理过程时间较短,具有一定的实用性.
In view of the severe degeneration of underwater image and difficulty in image information extraction,an underwater degenerative image processing method was presented.With the analysis of underwater image degeneration process,the method of obtaining underwater image degradation function based on turbulence model was proposed,and frequency domain filter was adopted for underwater image restoration.An image segmentation algorithm using 2D maximum entropy threshold was presented to process meaningful segmentation of images,which combined artificial fish swarm algorithm(AFSA)with image 2D Abutaleb entropy information.As AFSA only compares the advantages and disadvantages of problems without obtaining their special information,the searching process of AFSA is fast and adaptive.Experimental results have verified the feasibility of the proposed method,which can deal with the problem of fuzziness and low contrast of under water degenerative image effectively and achieve better segmentation results with less time.
出处
《天津大学学报》
EI
CAS
CSCD
北大核心
2010年第9期827-833,共7页
Journal of Tianjin University(Science and Technology)
基金
国家高技术研究发展计划(863计划)资助项目(2008AA092301)
国防科技重点实验室开放课题研究基金资助项目(2008003)
关键词
水下图像复原
水下图像分割
退化函数
湍流模型
人工鱼群算法
二维熵阈值
underwater image restoration
underwater image segmentation
degradation function
turbulence model
artificial fish-swarm algorithm
2D entropy threshold