基于二维高斯曲面拟合的图像灰度补偿算法

A compensation algorithm of image based on Gaussian quadrics fitting

研究了基于CCD的带钢表面缺陷检测系统图像采集中出现的灰度失真问题 ,提出了在不改变系统硬件配置的基础上 ,根据主导退化因素采用二维高斯曲面为基本退化模型 ,同时兼顾其他没有明确数学模型的图像退化因素 ,根据图像灰度分布统计特性估计二维高斯曲面各参数以实现图像灰度补偿的软解决方法 ,实验结果表明 ,该算法有效地提高了图像质量 。

The restoration of image collected in strip steel ins pe ction system was studied based on CCD. A Gaussian quadrics was presented as a ba sic model of the main restoration caused by the departure from optic focus of CC D. Other causes were considered without specific mathematical models. The parame ters of Gaussian quadrics were estimated according to the statistical distributi on information of gray image. The result revealed that Gaussian fitting was a go od solution to complex restoration problems.