摘要
为了解决图像恢复时所引起的阶梯效应和边缘模糊问题,定义可变TV_p范数,提出一个自适应TV_p(Adaptive TV_p,ATV_p)正则恢复模型,并结合AOS数值计算方法,给出完整的ATV_p正则恢复算法,其中p可以自动区分图像中的边缘和平坦区域,自适应选择不同的数值,使得新模型在恢复的同时不仅能够自适应的对图像中目标边缘进行有效的保护,而且可以避免出现阶梯效应。实验表明,和主要的一些正则模型相比,本恢复算法对模糊图像的恢复无论在视角效果还是定量指标上都有了明显的改进。
In order to avoid the staircasing effect and edge blurring problem. A variable TVp norm was defined, and an adaptive TVp (ATVp) regularization model was proposed. Combining the AOS numerical method, a complete ATVp regularization algorithm was shown, where p can be adaptive selected according to different image areas. The characteristics make the new model preserve the edge information better and avoid the staircasing effect while image restoration. Experiments showed that compared with the existing regularization models, it improved the restoration results in both visual effects and SBR and PSNR.
出处
《中山大学学报(自然科学版)》
CAS
CSCD
北大核心
2016年第5期8-13,共6页
Acta Scientiarum Naturalium Universitatis Sunyatseni
基金
国家自然科学基金面上资助项目(11271381
11471339)
国家自然科学基金青年基金资助项目(61301229)
河南省教育厅资助项目(15A110020)
河南科技大学博士科研启动基金资助项目(13480032)
广州市科技计划资助项目(201607010144)