摘要
采用改进模板匹配和最小二乘法对偏心桶加注口进行视觉定位。首先通过阈值分割和中值滤波对桶口图像进行预处理,然后通过改进的相关系数法进行模板匹配,获得要识别区域,减小运算量。最后通过最小二乘法得到偏心桶加注口的圆心坐标。实验结果表明,该方法的精度能达到工业要求,相比传统Hough变换,实现了检测的实时性。
This paper uses improved template matching and the least square method to realize visual positioning of the eccentric bucket filler. First of all, threshold segmentation and median filter are used to preproeess the barrel mouth image, and then template matching is carried out by improved correlation coefficient method to get identify regions and reduce computation. Finally, the center coordinates of eccentric bucket filler is obtained by least square method. The experimental resuhs show that the precision of this method can meet the requirement of industry. Compared with the traditional Hough transform, the method can improve the real-time detection.
出处
《机械工程师》
2015年第2期73-75,共3页
Mechanical Engineer
关键词
偏心桶加注口
阈值分割
中值滤波
模板匹配
最小二乘法
eccentric bucket filler
threshold segmentation
median filter
template matching
least square method
