AOI检测系统光照不均的校正方法研究

针对电子组装业自动光学检测(automatic optical inspection,AOI)系统光照不均情况,分析了背景去除法、传统同态滤波法2种校正方法,并将一种改进的基于小波变换的同态滤波方法应用于该领域。该方法对目标图像进行多层小波分解,对各层高低频系数滤波处理,减弱低频系数,增强高频系数,然后进行小波重构,从而达到光照不均校正的目的。然后取PCBA(printed circuit board assembly)图像进行了3种方法的实验对比,结果证明了改进的基于小波变换的同态滤波方法在AOI系统中具有很强的实用性。

Research of illumination correction in AOI system

In this paper,two methods for illumination correction in AOI system are analyzed,which are de-background and conventional homomorphic filtering.A new analysis method of homomorphic filtering based on wavelet is introduced.The wavelet analysis coefficients are processed using a highpass filter to attenuate the low spatial frequencies and amplify the high spatial frequency.So the undesired contributions within the image due to light source ununiformity can be reduced and the object features can be emphasized.Experiment results show that the new analysis method of homomorphic filtering based on wavelet method performs well in illumination correction.