摘要
在相机曝光时间内,相机与拍摄目标存在相对运动会造成图像运动模糊。针对运动模糊图像,在辨识运动模糊参数的基础上,提出了一种逐行法来恢复运动模糊图像。运动方向通过运动模糊图像的频谱和Radon变换得到,利用得到的运动方向将频谱图旋转至水平方向。对处理后频谱图上的像素按列累加到水平方向上得到其投影图,在投影图上搜索对应暗条纹的极小值点,通过一系列极小值点的间距计算运动模糊尺度。最后,采用逐行法来恢复运动模糊图像。实验结果表明:算法辨识的运动模糊参数精确,辨识模糊尺度的误差在一个像素以内;同时逐行法恢复图像效果优异,可获得细节清晰的图像。
In the camera exposure time, the image motion blur is caused by relative motion between the camera and the objective. For the motion blurred image, a novel line-by-line method to restore the blurred image is pro- posed on the basis of identification of motion blur parameter. The motion direction is gained by Fourier spectrum of motion blurred image and Radon transformation. And then the motion blur direction is rotated to horizontal direc- tion. Pixels in the Fourier spectrum of motion blurred image is added to the horizontal direction by the column, then its projection is obtained. The minimum points corresponding were searched to the dark stripes in the projec- tion. Motion blur length is calculated by the distance between a series of minimum points. Finally, the line-by-line method was used to restore blurred image. The experimental results show that the algorithm has high accuracy for identification of motion blur parameter, the error of motion blur length is less than 1 pixel, at the same time, resto-ration based on line-by-line method has more better restore effect and gain much clearer details.
出处
《科学技术与工程》
北大核心
2016年第5期177-181,共5页
Science Technology and Engineering
基金
河北省自然科学基金项目(E2013209266)
河北省高等学校科学技术研究项目(ON2013114)资助
关键词
运动模糊
运动模糊图像频谱
RADON变换
投影
逐行法
motion blur line method spectrum of motion blurred image Radon transform projection line -by-
