调焦系统中数字图像清晰度评价函数的研究

摘要

In order to slove the problems of auto-focusing technology based on image processing, it is the key step to use the accurate and effective functions.The role of Gradient function, Spectrum function and Entropy function in auto-focusing system was analyzed and compared by experiment in auto-focusing system.The performances and features of these functions in experimenteral platform were studied.Experiments results indicate that Tenengrad function, Brenner function and Square Gradient function has high sensitivity features in focal plane,when defocus serious, the tread of change is not clear, this can be applied in accurate focus equipment.Variance function has good SinglePeak and sensitivity, characterize the preferable efficiency, has high single-to-noise, better reliability, than other ones.It can better meet the requirement of the image definition evaluation, can be applied in digital image auto-fccusing system.%为了实现基于数字图像处理的自动调焦技术,采用准确有效的图像清晰度评价函数是其实现的关键.通过实验分析比较了梯度函数、频谱函数和熵甬数在自动对焦技术中的作用,并研究了各评价函数在实验平台下所表现出来的性能和特点.研究结果表明,Tenengrad函数、Brenner函数和平方梯度函数在焦平面附近具有变化趋势明显和灵敏度高的特点,而在离焦较远时变化不大,适用于要求精确对焦的设备.而方差甬数具有良好的单峰性和无偏性,具有较好的计算效率,相对其他的评价方法具有更高的信噪比、可靠性和清晰灵敏度,更好地满足了对图像清晰度评价的要求,适用于摄像头辅助调焦系统中.