基于直线检测的运动模糊长度估算算法  

The Estimation Algorithm of Motion Blur-length Based on Line Detection

作　　者：谢晴[1] 玉振明[2] 

机构地区：[1]梧州学院计算机科学系,梧州543002 [2]梧州学院信号与信息处理实验室,梧州543002

出　　处：《科学技术与工程》2013年第8期2113-2117,2123,共6页Science Technology and Engineering

基　　金：国家自然科学基金资助项目(61062014);广西教育厅科研项目(201106LX571);梧州学院院级重点项目(2011B008)资助

摘　　要：理论上论证了匀速直线运动模糊图像的频谱图中存在等间隔的黑色条纹的现象。提出通过模板滤波直线检测方法识别频谱图中间亮区域的边界线,计算边界宽度。根据相邻黑色条纹间距与模糊长度的数量关系,最终估算出模糊长度的方法。对大量不同模糊程度的图像均基于MatLab7.0平台,进行了实验仿真。结果表明,与其他方法相比,所提算法能较准确的估算出模糊图像的模糊长度,为模糊图像复原打下基础。The spectrum of uniform linear motion-blurred image ted, presented use the line detection of template filter method for ide exists equi-spaced black stripes is demonstrantifying lines of bright area in the middle of the spectrum, calculating border width, estimate blur-length according to relationship of adjacent black stripe space and blur-length. Experiments using many different degree of blur-image based on matlab7.0 are conducted. Results indicate that the algorithm can estimate the blur-length more accurately and lay the foundation for motion-blurred image restoration.

关 键 词：直线检测 模糊长度 频谱图 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]