光学与光电技术, 2008, 6 (4): 0068, 网络出版: 2010-05-24   

基于频谱分析的匀速运动模糊图像模糊方向识别

Identification of Motion Blur Direction from Motion Blurred Image by Spectrum Analysis
作者单位
国防科技大学光电科学与工程学院,湖南 长沙 410073
摘要
点扩展函数的设置是图像复原中的关键问题,对于匀速直线运动模糊图像,运动方向和模糊长度决定了点扩展函数。根据运动模糊图像和原始图像在频谱上存在的对应关系,即运动模糊图像频谱存在着对应于传递函数零点的暗条纹,提出对运动模糊图像,通过二维傅里叶变换、二值化以及Radon变换检测运动模糊图像频谱图上暗条纹的方向和位置,来实现运动方向测量的方法。用该方法对模糊图像进行检测,模糊方向的识别精度小于1°。实验证明该方法可以实现平面内任意方向运动的测量。
Abstract
Setting the point spread function of motion blurred image is the key problem in the image restoration. For uniform linear motion blurred images, the blur direction and length determine the point spread function. The identification of blur direction in motion blurred images is one of the most important steps of image restoration. There are sets of negative ripples corresponding to zero values of the transfer function, which can be found on spectrums of motion blurred images. Based on this spectrum relationship between motion blurred image and original image, a method to identify the motion blur direction from the motion blurred image is presented by detecting the negative ripples direction and location on the 2D-Fourier transform spectrum with binary image and Radon transform. The method is used to detect motion blurred images and vague direction. The identification accuracy is less than 1°. The experimental results prove that this method is able to measure motion blur in discretionary direction.

蒋岩峰, 于起峰, 梁永辉. 基于频谱分析的匀速运动模糊图像模糊方向识别[J]. 光学与光电技术, 2008, 6(4): 0068. JIANG Yan-feng, YU Qi-feng, LIANG Yong-hui. Identification of Motion Blur Direction from Motion Blurred Image by Spectrum Analysis[J]. OPTICS & OPTOELECTRONIC TECHNOLOGY, 2008, 6(4): 0068.

本文已被 3 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!