光学测量中常用的数字散斑相关方法(DSCM)目前缺少有效的对匹配效果的评价方法,尤其是没有对匹配精度的估计方法。针对这种情况,提出了一种评价指标体系。该体系包含六个参数,其中,简单信噪比越大主峰被漏检的可能越小;主峰底部截面面积越大越容易搜索到主峰;主峰平滑系数与匹配精度有显著的相关关系,适中的主峰准峰顶截面面积和较大的主峰平滑系数有利于获得较高的匹配精度。灵活运用六个评价指标,可利用小范围实验数据辅助确定相关测度函数、搜索算法、亚像素插值方法等,最大限度地减少低效率匹配运算,优化匹配质量。选取了一个确定模板尺度的实例验证了该方法客观有效。
As a commonly used method in optical measurement, digital speckle correlation method (DSCM) lacks of effective evaluation methodology for registration effects, particularly for registration accuracy. So this paper proposed an evaluation sys- tem including 6 parameters. A greater SSNR ( simple signal-to-noise ratio) reduces the possibility of missing main peak. A greater SL(area of the bottom of the main peak) means a easier search for the main peak. F( main peak smoothness factor) and registration error has significant negative correlation. A moderate St( area of the top of main peak) and a greater F is con- ductive to a higher degree of registration accuracy. Flexible usage of the 6 parameters can aid to choose correlation measure function, determine search algorithm, abd and to determine sub-pixel interpolation method, etc. An example of determining the template' s scale demonstrates the validity and objectivity of the system.