Hough 变换在电路板检测中的应用

针对 DSP56F801 最小系统电路板检测系统中每次采集到的电路板图像存在平移和旋转误差的问题,提出了基于 Hough 变换的电路板图像矫正算法.该算法首先利用边缘检测算子得到电路板图像的边缘图像,再根据该电路板中定位孔位置固定的特点,利用 Hough 变换检测电路板定位孔的圆形外边缘的位置,然后以定位孔外边缘的圆心为基础计算出电路板的平移误差和旋转误差并进行矫正.实验验证了这一改进的效果.消除了该检测系统中电路板图像的平移误差和旋转误差.

Application of Hough transform in detection of circuit board

The problem of the image of the circuit board everytime collected from the auto-detection system exist parallel motion and circumrotatory warp was researched. This auto-detection system aim at the minimum system circuit board which model is DSP56F801. The rectificatory algorithm used in the image of the circuit board based on the hough transform was advanced. Firstly, in this algorithm the edge detect arithmetic operators was used to get the edge image of the image of circuit board, aim at the characteristic of the circuit board which location of the orientation holes is fixed, hough transform was used to detect the location of circular outer edge of the orientation holes of the circuit board. Then the parallel locomotor warp and circumrotatory warp of the circuit board can be computed and rectified based on the centre of the outer edge of the orientation holes. Finally, the algorithm was validated use the image of the circuit board actual collected. The experimental results indicate that this algorithm can eliminate the parallel locomotor warp and circumrotatory warp of image of the circuit board collected from the auto-detection system.