复杂背景下圆孔半径及圆心的图像检测方法

Approach to Detect Radius and Center of Circle from Complex Background

针对刷子插毛生产线的现实问题,本文给出了一种在复杂背景下圆孔的半径和圆心的图像检测方法。首先,根据扫描过程中角度不绝对垂直以及背景木板条纹大小和方向的不确定性,采用图像增强方法提高了对比度;然后,用边缘点检测方法来确定目标范围,用数学形态学的方法提取目标;最后,用平均误差的算法提高精度。使用本方法,可以用图像检测代替用手工标尺的测量。实验结果表明:相对于传统检测方法,本文的方法定位效果好,提取目标效果佳,具有进一步提高目标检测精度的潜力。

In order to find a solution to insert feather into brush in product line,an approach is proposed in this paper to detect radius and center of circle from complex background. Firstly,to overcome the problem that the angle is not absolute uprightness in the process of scanning and the size and direction of wood texture are uncertain, we adopt image amplification technique to improve contrast between target and background. Secondly,locate the district of target by using edge detection method,and detect it by mathematical morphology method. Finally, use the average error method to improve accuracy. By the proposed method, we can perform detection by image recognition instead of manual measure. Experimental results show that, relative to traditional detecting method,the proposed algorithm can detect target effectively and has potential to improve detecting precision.