摘要
针对缝合针尺寸微小针尖磨削质量难以检测的问题,提出了一种利用机器视觉技术提取缝合针图像边缘直线并根据直线参数进行磨削质量检测的方法。首先在图像预处理的基础上通过对比法选用Canny算子作为图像边缘检测算子,进一步分析了传统Hough变换直线检测原理,根据参数空间中θ参数与边缘梯度方向相同的特点,将边缘梯度方向引入Hough变换,缩小了参数θ的取值区间,提高了算法速度。最后给出了磨削质量的判定条件及算法实现步骤。针对0.5 mm的一组缝合针针尖检测实验结果表明,本文算法对针尖角度的检测与传统检测方法相比平均误差不超过0.53°,平均检测时间527.7 ms,工作状态下对针尖磨削质量判定的平均准确率达92%,实时性较好,具有一定的实用价值。
On the base of the difficult to detect quality of the grinding of suture needle tip with the small size of the suture needle, this paper presents a method of detecting the grinding quality with edge detection method based on machine vision technology. First, canny operator is chosen as the image edge detection based on image pretreatment after comparison, this paper also analyses the principle of a traditional way of line detection based on Hough Transform, according parameter θ and edge gradient is in the same direction, introduced Hough Transform into edge gradient, parameter θ is narrowed down, improves the arithmetic. Finally gives the distinguish condition of grinding quality and the realization of the algorithm. Experimental results of φ0.5mm show that the average deviation of this algorithm is lower than 0.53° with that of traditional way, the regular time is 527.7ms, the mean correct rate of the quality of the point of a needle grinding is 92% in normal working order, this method has a good real-time and certain practical value.
出处
《机械设计与研究》
CSCD
北大核心
2014年第5期139-142,共4页
Machine Design And Research
基金
淮安市科技支撑计划(工业)资助项目(HGZ2011016)
关键词
带线缝合针
机器视觉
边缘检测
Hougn变换
针尖磨削
suture needle with line
machine vision
edge detection
hough transform
grinding of suture needle tip
