基于Hausdorff距离的工业机器人视觉研究

Research on industrial robot vision based on Hausdorff distance

选用中值滤波的方式对原始的图像素材进行修饰,采用自动阈值的Canny算子获取工件的边缘特征信息。这些边缘特征组成了一块块的区域,用人眼观察,很容易找到目标区域,但是计算机很难找到目标区域。利用遗传算法从一块块的区域里匹配到目标工件,并对匹配的结果进行判定。文章采用改进的Hausdorff的方法对结果进行判定,减少计算量的同时提高判定结果的可信度。

The original image material is modified by median filtering,and the Canny operator with automatic threshold is used to obtain the edge feature information of the workpiece.These edge features form a block of areas.It is easy to find the target area with human eyes,but it is difficult for computers to find the target area.The genetic algorithm is used to match the target workpiece from a block of regions,and the matching result is judged.The improved Hausdorff method is used to judge the result,which reduces the calculation and improves the reliability of the judgment result.