球栅阵列芯片图像边缘提取方法研究  

Study on the Image Edge Extraction Method of Ball Grid Array Chip

作　　者：李克新[1] 王金聪[1] 宋文龙[1] 

机构地区：[1]东北林业大学机电工程学院,哈尔滨150040

出　　处：《森林工程》2013年第3期87-89,共3页Forest Engineering

基　　金：黑龙江省科技计划项目(GC06A522)

摘　　要：针对球栅陈列(BGA)芯片图像的特点,其边缘提取最为关键技术的是芯片的引脚图像的轮廓提取。一般传统的边缘检测方法大都是以原图像为基础,检测边缘是利用边缘临近一阶或二阶导数方向变化规律,对噪声极度敏感,因此常常边缘误检。通过蛇模型及其改进处理BGA芯片边缘检测,使其对噪声的敏感程度尽量按要求降到最低,同时在不漏测真边缘的前提下尽量不引入假边缘,运算速度快,效果好,抗干扰性强。For the characteristics of the image of BGA chip, the contour extraction of chip pin image is the most critical technolo- gy in its edge extraction. By comparing to the traditional operator, the edge extraction algorithm based on the snake model was used in this paper. The traditional edge detection method is mostly based on the original image, using edge near the first or second order de- rivative of the direction of changes in the law to detect edges, which is extremely sensitive to noise, so it often cause the false alarms of edge detection. To this end, the paper handles with the edge detection of the BGA chip through the snake model and its improved, which overcomes the sensitivity to noise to minimum and does not introduce false edge in the premise of no leaking the true edge. The improved model can speeds up the computing speed, brings good effect, and has strong anti-interference.

关 键 词：BGA芯片 边缘提取 图像处理 蛇模型 

分 类 号：S776.01[农业科学—森林工程]