元件疵病检测中照度不均匀图像的融合处理方法

Combined Processing Method of Uneven Illumination Images for Defect Detection of Components

针对光学元件疵病检测中,通过光学成像获得的图像照度不均匀,噪声明显的问题,提出了根据元件检测照明区的分布特征进行融合处理的方法;该方法采用开运算提取背景算法和高斯差分滤波器,开运算提取背景算法可以给出没有直接照明区域的疵病分布,高斯差分滤波器可去除照明光源的噪声影响;实验结果表明,相较于单一的处理方法,所提出方法更好地得到了整个区域内的疵病分布,有效地实现了不均匀照度环境下的疵病分辨。

In the defect detection of the optical components,the images got from optical system have noise area and uneven illumination background.A combined processing method based on the feature of illumination distribution is proposed.The method utilizes open operation and difference of Gaussian filter（DOG）.By open operation,the background is extracted and defect distribution of the dark area is acquired then.By DOG,the influence of the light source noise is eliminated.The experiment results show that comparing with single method,the combined method lets the defect distribution of the optical component better emerge.It is proved that the suggested method is effective in obtaining the exact defect distribution from images taken in the condition of uneven illumination.