论文部分内容阅读
在瓷砖视觉检测中,提出了一种新的定位分割方法,该方法可对任意角度的瓷砖目标进行定位分割,为后续的缺陷检测及机器人抓取提供了高精度的处理模块与定位基准。首先通过预处理消除采集图像噪声,然后根据瓷砖自身的灰阶梯度,对其进行轮廓提取,并以轮廓斜外接矩形面积及矩形度双重约束去除复杂背景的干扰。再通过轮廓掩膜获取目标区域图像,设置ROI(Regions of Interest,感兴趣区域)使目标区域从检测背景中分离出来成为一个独立单元。最后采用斜矩形对角线与标准位置的几何偏差作为矫正的反馈信息,以矫正轮廓质心作为定位基准。
In tile visual inspection, a new positioning segmentation method is proposed, which can locate and segment tile targets at any angle and provide high-precision processing modules and positioning datum for subsequent defect detection and robotic grabbing. Firstly, the image noise is removed by preprocessing, and then the contour of the tile is extracted according to the grayscale gradient of the tile itself. The complex background interference is removed by the double constraint of circumscribed rectangle area and contours. And then through the contour mask to obtain the image of the target area, set the ROI (Regions of Interest, region of interest) so that the target area from the detection background to become an independent unit. Finally, the geometric deviation of the diagonal rectangular diagonal and the standard position is taken as the feedback information of the correction to correct the contour centroid as the positioning datum.