摘要
建立边缘对准度函数,通过抽样特征点对样本PCB图像与待测PCB图像进行全局配准.由于相机镜头畸变导致图像非线性失真,全局对准后,局部配准存在偏差.提出分块对准策略,解决局部配准问题,同时针对局部配准后,块与块的交界处存在断裂现象,采用最小二乘法进行边缘拟合,实现全局PCB与局部PCB图像无缝配准.
The paper established the edge alignment function firstly and carries on the global registration for the sample PCB image and testing PCB image through the sampling characteristics point. Because camera lens distortion caused the image non-linear distortion, the local registration had some deviations after the overall registration. We proposed block registration strategy, to solve local image alignment issue and simultaneously carried on the edge fitting with the method of least square according to the ruptured phenomenon between the block and block. Finally image seamless matching was realized for both whole PCB and local PCB image.
出处
《安徽大学学报(自然科学版)》
CAS
北大核心
2013年第6期13-17,共5页
Journal of Anhui University(Natural Science Edition)
基金
国家自然科学基金资助项目(61272278)
湖北省教育厅自然科学基金资助项目(B2013222)