提出一种视觉机器人货物拣选中的货物轮廓提取与定位方法.在图像预处理的基础上,采用sobel算子提取货物与货箱边缘图像I1[i,j],通过图像投影取得货物支持区及其边缘图像I2[i,j],对货物支持区利用Canny算子提取边缘图像I3[i,j],利用I3[i,j]对I2[i,j]进行边缘连接,最后边缘跟踪提取货物轮廓实现货物定位,控制机械手运动并实现货物抓取.实验结果表明该方法实用有效.
A new method of goods contour detection and positioning by a vision robot was proposed. Based on image pre-processing and the Sobel operator, the edge of goods and its container I1 [ i, j ] was detected. The goods-supporting area and its edge image I2 [ i,j] were calculated'by image projection. Then another goods-supporting area's edge image I3 [ i,j] was obtained according to the Canny operator and the breakpoints in I2 [ i ,j] were connected by comparing with I3 [ i,j]. Finally, the contour was extracted by the contour extraction method to locate the target and the robot was controlled to pick the goods. The experiments show that the method is effective and reliable.