摘要
为了更精确地检测出含有噪声与模糊的印刷电路板(PCB)光电图像中走线的信息,提出了结合随机Hough变换和最小二乘法拟合直线两者优点的直线检测新方法。首先对含有走线的PCB光电图像进行预处理,给出误差阈值;然后用随机Hough变换粗略确定图像中直线的大概位置,剔除干扰点与噪声,保留符合误差阈值要求的直线附近的点集;最后用最小二乘法拟合各个集中的点,得到较精确的直线参量,从而得到各条拟合直线。对由CCD与显微镜获取PCB光电图像的走线进行实际检测,结果表明,本文提出的方法可实现精确的直线检测。
In order to accurately detect the lines in the printed circuit board（PCB） photoelectric image with noise and illegibility,a novel method combining the random Hough transform with the least square fitting is proposed and realized.Firstly,the preprocessing of PCB photoelectric image with lines is made,and the error threshold is given.Secondly,the approximate positions of lines in the detected image are roughly by random Hough transform,the disturbance points and noise far away from the lines are deleted,the point sets that satisfy the request of error threshold in the vicinity of each line are reserved.Lastly,the accurate parameters of lines are obtained by fitting each point within each set by the least square method,and then each line in the PCB image is obtained.The experiments for detecting the lines in PCB photoelectric images acquired by CCD and microscope are executed.The results verify that the lines in PCB can be accurately detected by the proposed method.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2010年第12期1885-1889,共5页
Journal of Optoelectronics·Laser
基金
总装基金资助项目(9140A01020507DZ02)
国家重点实验室基金资助项目(9140C1406020708)
粤港关键领域重点突破资助项目(20091683)
湖南省科技计划资助项目(2009FJ3187)
