摘要
模糊是图像退化的主要降质因素之一,图像去模糊技术一直是计算机视觉和图像处理领域的一个重要课题,广泛应用于安防、刑侦、交通、金融、医疗图像等领域.首先,对传统去模糊算法进行阐述并说明其存在的问题.然后,在此基础上对现有基于深度学习的图像去模糊方法进行综述并分析了算法存在的优势与不足,着重讨论了基于动态网络的非均匀模糊处理新趋势.最后,结合最新的研究成果,展望了图像去模糊技术未来的发展方向.
Blur is one of the main factors of image degradation.Image deblurring has always been an important subject in the field of computer vision and image processing,and is widely used in security,criminal investigation,transportation,finance,medical imaging and other fields.First,the traditional deblurring algorithm and its problems were explained.Then,on the basis of this,the existing deep learning-based image deblurring methods were reviewed and the advantages and disadvantages of the algorithms were analyzed.Finally,combined with the latest research results,the future development direction of image deblurring was proposed.
作者
李晓光
杨飞璠
卓力
LI Xiaoguang;YANG Feifan;ZHUO Li(Faculty of Information Technology,Beijing University of Technology,Beijing 100124,China)
出处
《北京工业大学学报》
CAS
CSCD
北大核心
2021年第8期982-990,共9页
Journal of Beijing University of Technology
基金
国家自然科学基金资助项目(61701011)
北京市自然科学基金资助项目(KZ201810005002)。