面向自动对焦的图像预处理技术

从噪声、亮度一致性及视场变化等角度分析了在对焦系统中进行图像预处理的必要性。提出维纳滤波与亮度归一化相结合的图像预处理技术,去除固定模式噪声和亮度变化的影响。实验表明,经过预处理后明显改善了对焦评价函数曲线的形状:曲线更加平滑,局部峰值得以去除,同时曲线灵敏度提高78%。所提出的预处理技术使对焦搜索避免陷入局部极值,在提高对焦速度与对焦精度两方面均有应用价值。

Image pre-processing techniques for auto focusing

The necessity of image pre-processing in focusing system is analyzed from the aspects of noise, consistency of brightness and field of view. An image pre-processing techniques based on Wiener filter and normalization of image brightness is proposed for weaken the influence of periodic noise and brightness variation. The experiments show that the shape of focusing measure curve is obviously improved after pre-processing, the curve is smoother, local peak is eliminated and curve sensitivity is improved by 78%. The proposed pre-processing techniques makes the focusing search escapes influence of local peak and it has application value both in optimization of focusing speed and focusing accuracy.