摘要
提出了一种针对不同形状工件的视觉检测识别方案,通过采用中值滤波等算法对图像进行预处理,采用基于Hu矩和SVM分类器的机器学习等方法完成对不同形状工件的识别及相关尺寸参数计算。实验结果表明,此检测识别方案可有效地识别出一幅图像中的不同形状工件,并计算出工件相关特征的尺寸参数。
This paper proposed a solution, which is used for detection and recognition workpieces in different shapes with computer vision technology. Firstly, median filtering is used for noise filtering during image pre-process- ing stage. And then, machine learning based on Hu moments and SVM classifier is applied to identify different ob- jects. Experimental results showed that, this solution can detect and identify objects in different shapes effectively and successfully, and it can also calculate characteristics of these objects such as position coordinates and dimensions in pixels.
出处
《机电一体化》
2016年第2期8-11,30,共5页
Mechatronics
