摘要
在日益自动化的今天,必须快速有效地实现图像测量的自动化.而要达到此目的,首先必须选择一个合适的图像自动聚焦评价函数.在研究了众多图像聚焦评价函数的基础上,发现梯度算子存在一些不足之处,为此将阈值引入到该算子中.经过大量的仿真实验,结果表明,引入阈值的算子的聚焦性能有了明显的改善,而且比其他的评价函数也更加准确、稳定、可靠.
With the automatization increasing, the rapid and efficient auto-focusing in the measurement of images should be achieved. In order to realize this, an appropriate auto-focusing evaluation function for images should be chosen at frrstly. Based on the auto-focusing algorithms investigation, it is found that there are some flaws in the auto-focusing evaluation function of the Roberts gradient. So a threshold parameter into the algorithm is introduced. Based on careful experiments, it is found out that the Roberts algorithms with the parameter have evident improvements in accuracy, stability and reliability.
出处
《山东大学学报（理学版）》
CAS
CSCD
北大核心
2006年第5期119-123,共5页
Journal of Shandong University(Natural Science)
基金
山东省自然科学基金资助项目(Y2005G08)
关键词
自动聚焦
评价函数
阈值
梯度
auto-focusing
the evaluation function
threshold
the roberts gradient
