一种改进的球栅阵列封装焊点射线图像阈值分割算法

An Improved Thresholding Segmentation Algorithm of Ball Grid Array X-ray Digital Image

球栅阵列封装焊点的射线图像具有信噪比差、背景不均匀等特点,故传统的阈值分割方法无法将目标焊点与背景图像很好的分割.本文通过对球栅阵列封装焊点射线图像直方图的分析,利用了自适应维纳滤波对阈值分割前的图像进行了预处理.根据图像的差异来调整该滤波器的参量,对局部差异大的地方进行小的平滑操作,对局部差异小的地方进行大的平滑操作.在最大类间方差法的基础上,对分割后的图像进行了进一步的分析并提出了改进的二次分割方法.改进的方法为并不直接通过OTSU法进行二值化处理来去除背景,而是在阈值分割得到的两个灰度级内通过计算中值和统计最大灰度像素的方法得到了更优化的阈值,使得去除背景后的焊点图像整体更加清晰和均匀.在背景灰度级内寻找了一个合适的灰度级作为处理后的灰度图像新背景,实验证明该方法明显改进了传统最大类间方差法对球栅阵列封装焊点射线图像的阈值分割效果.

The X-ray image of BGA has low SNR characteristic and nonuniformity characteristic.Traditional threshold segmentation is unable to segment BGA solder from background accurately.By analyzing of the histogram of the BGA X-ray image,Wiener filtering was used in the image preprocessing procedure.The filter parameter was adjusted for different images.Low and high smooth operations were used in minor and bigger differences,respectively.OTSU method was implemented to select a initial threshold,and an improved threshold segmentation algorithm was used to segmented images based OTSU method.It meaned that the BGA image would not be processed directly by OTSU method.In fact,a better threshold would be found by calculating the mean value and calculating the max statistic of pixels.A new pixel value was chosen to replace all pixels value of the background in BGA image.The experimental results show that this algorithm improves the segmented effect compared with the traditional OTSU method.