基于光流约束和光谱蒙板的空间变换运动模糊图像复原  

Spatially-varying Blurred Image Restoration Approach Based on Motion Blur Constraint and Spectral Matting

作　　者：杨斌[1] Yang Bin(Guangzhou Vocational College of Technology & Business,Guangzhou 511442,China)

机构地区：[1]广州科技贸易职业学院,广州511442

出　　处：《科技通报》2016年第12期155-158,168,共5页Bulletin of Science and Technology

基　　金：基金项目:基于移动终端的听评课互动平台的开发与应用(编号:20151317111)

摘　　要：针对复杂运动模式的模糊图像,提出了基于光流约束和光谱蒙板的空间变换运动模糊图像复原方法。该方法运用运动模糊约束条件和alpha蒙板简化了图像复原模型,采用alpha通道模型减少了自由度和复杂度。本文方法不需要用户的干预和模糊方向的假设等先验信息,复原性能和效果优于当前技术条件下的空间变化运动模糊复原方法。该方法能够很好地完成了空间变化运动模糊图像的复原,实验结果表明本文方法的有效性和实用性。Aiming at blurred image of complex mode motion, an spatially- varying blurred image restoration approach based on motion blur constraint and spectral matting was proposed. This method uses the motion blur constraints and alpha matting simplifies the image restoration model. Alpha channel model is adopted to reduce the degree of freedom and complexity. The user's intervention and the assumption of fuzzy direction is not needed. The recovery performance and effect is better than that of the current technology. The method can accomplish the restoration of the blurred image of the spatial variation. Experimental results show that the proposed method is effective and practical.

关 键 词：运动模糊 图像复原 模糊图像 

分 类 号：TP273.5[自动化与计算机技术—检测技术与自动化装置]