机器人视觉中障碍物刚柔性的判别方法

Obstacles rigidity and flexibility discriminant method in robot vision

为了降低机器人避障算法的复杂度、提高避障效率,提出了一种在避障前对检测到的障碍物的刚柔性进行判定,从而除去不需要进行避障处理的柔性物体的方法。具体方法是将提取出的视频图像的帧图序列进行处理后叠加,提取叠加图像外边缘的特征点,变换坐标系后进行曲线拟合,得到特征曲线的方程式,根据其计算出障碍物的带权柔度,将障碍物的带权柔度和柔度阈值作比较,当障碍物的带权柔度大于或等于柔度阈值时,障碍物为柔性物体;否则,障碍物为刚性物体。此外,提出了一种物体刚柔性的误判情况和解决方法以及在无风或微风情况下该方法局限性的解决方案。实验结果证明,这种方法可以很好地判别物体的刚柔性。

In order to reduce the complexity of the robot obstacle avoidance algorithm and improve the efficiency of obstacle avoid- ance, this paper proposes a method of distinguishing the rigidity and flexibility of obstacles before obstacle avoidance, and remov- ing flexible objects which needn＇t obstacle avoidance. The specific method is to process and overlay the extracted video image frame sequence, and then extract the feature points of the superimposed image outer edge, fit the curve after transforming the coor- dinate system, get the equation of the characteristic curve and calculate the obstacle＇ s weighted flexibility in accordance with it, then take the obstacle＇ s weighted flexibility and flexibility threshold for comparison; when the weighted flexibility of the obstacle is greater than or equal to the flexibility threshold, the obstacle is flexible; otherwise, the obstacle is rigid. In addition, it proposes an object rigidity and flexibility miscarriage justice as well as the solution and the solution to the limitations of the method proposed in the paper in no wind or breeze. Experimental results show that this method can well distinguish the object of rigidity and flexibility.