一种改进的数字图像清晰度评价函数研究

图像自动聚焦时，由于噪声和不同光照条件影响图像质量，实时评价图像清晰度效果较差。本文比较了能量梯度函数、频谱函数、灰度熵函数、灰度变化函数评价图像清晰度的优劣，提出了一种改进的能量梯度闽值方法评价图像清晰度，建立了数学模型，并给出了三种不同物体的数字图像聚焦重复精度的试验结果和分析。试验结果表明：实时评价图像清晰度该方法速度快，计算量小，信噪比高，可靠性好，聚焦重复精度高。

Research of evaluation function of an improving digital image definition

Because noises and different illumination conditions for auto-focus can affect image quality, the effect is worse for realtime evaluation of image definition. In the paper it is compared to use energy gradient, Frequency, Gray Entropy and Gray Level Variance as evaluation functions for their advantages and disadvantages, it is proposed to take advantage of energy gradient function threshold to evaluate image definition, and the mathematical model is established. Furthermore, the experiment results and analysis of the precision of repetition of focus are displayed for three different subjects. The experiment results indicate for real-time evaluation of image definition this method has faster speed, less account quantity, high Signal-to-Noise, better reliability, high precision of repetition of focus.