摘要
针对目前清晰度检测方法中存在的速度慢、局部量化不明显和受光照亮度影响大等不足,在分析了几种常见图像清晰度评价函数特性的基础上,提出了一种基于图像中心距的清晰度评价模型。该模型通过引入图像中心距和高通滤波图像中心距的线性组合,在保持评价结果良好全局性的同时提高了其局部量化精度。在此基础上实现了一种基于五区域的图像清晰度评价方法。实践表明该方法具有较好的计算效率、可靠性和量化等级。
There are some deficiencies in actual methods of detecting image sharpness, such as low speed, unconspicuous local quantization and great influence ofillumination intensity. In order to overcome these deficiencies, after analyzed characters ofseveral popular functions in image Sharpness evaluation, a new image sharpness evaluation model was presented which is based on the image absolute central moment. The linear combination of image absolute central moment and high pass image absolute central moment was introduced into this model to increase the precision of local quantization as well as keep nicer entireness in evaluation results. On this basis, a estimating method of image sharpness based on five regions was put forward. Experiment shows that the method is more efficient, reliable and it＇s quantization is more preferable.
出处
《计算机工程与设计》
CSCD
北大核心
2006年第9期1545-1546,1597,共3页
Computer Engineering and Design
基金
江苏省研究生创新基金项目(xm04-35)
关键词
数字图像
清晰度
图像中心距
评价方法
digital image
sharpness
image central moment
estimating method
