摘要
多引脚小间距的IC元件是电子产品生产过程中视觉检测与定位的难点。针对传统的图像匹配算法的操作复杂,以及基于IC元件背面特征等定位算法的计算量大,适用性差等缺陷,提出了一种改进的差分直线检测的图像定位算法,该算法利用先验知识确定IC元件黑色塑基间边缘的大概区域,在这个大概区域里面采用一阶差分运算,提取黑色塑基的边缘点集合。该算法不需要对图像中的每一个像素点进行计算,因此大大地提高了算法速度,同时通过此方法提取的边缘点图像噪声点极少,使得后续的直线检测速度也得到了提升,实验结果表明,该算法的精度能满足实际生产需要,具有更好的适用性,为进一步检测芯片缺陷提供了保障。
It's difficult for multi-pin small pitch IC devices to be visually detected and located in the production process of electronic products.Due to the complicated operation of the traditional image matching algorithm,large computation and poor applicability of the positioning algorithm based on the back features of IC components.an improved image localization algorithm for differential line detection method is proposed.First,the priori knowledge is used to obtain the approximate area of the edge between the black plastic base of the IC component,then,by utilizing the first-order difference operation,the edge set of the black plastic base can be extracted.The speed of the method can be dramatically improved because it's no need to calculate every pixel in the image,namely,the edge point image with few noise can be extracted by the method,which makes the subsequent line detection speed upgraded.The experimental results show that the accuracy of the algorithm can not only meet the actual production needs,but also has better applicability,which provides a guarantee for further detection of defective chip.
作者
殷政
何周浩
周巨
黄月琴
罗兵
Yin Zheng;He Zhou-hao;Zhou Ju;Huang Yue-qin;Luo Bing(Intelligent Manufacturing Department of Wuyi University,Guangdong Jiangmen 529020)
出处
《电子质量》
2019年第1期1-4,8,共5页
Electronics Quality
关键词
IC元件
视觉检测
定位算法
IC component
Visual inspection
Localization Algorithm