基于机器视觉的Delta机器人分拣系统算法

Sorting System Algorithms Based on Machine Vision for Delta Robot

  • 摘要: 针对分拣过程中视觉系统对工件的重复拍摄问题,提出一种基于时间与工件位置的图像去重复算法,以实时分拣系统的系统运行时刻作为各单元的时间基准,将预测的工件到达某一固定参考位置的时刻与工件当前位置组合成一组能唯一识别工件的坐标,经周期性比较,判断并去掉重复图像信息.同时为提高分拣效率,提出一种基于牛顿-拉夫森迭代的动态抓取算法,建立了机器人跟踪工件的数学模型,并通过牛顿-拉夫森方法求解该非线性数学模型.最后用 MATLAB 对动态抓取算法进行了验证.样机实验中最快分拣速度达 110 次/min,误抓率小于 2‰,漏抓率为 0,证明了算法能够满足实时性要求,同时具有较高的准确性和稳定性.

     

    Abstract: To overcome the repeat shooting to workpieces by vision system in sorting process, an image deduplication algorithm based on time and workpieces' positions is proposed. The running time of the real-time sorting system is used as basis of each sorting module, and the predicted time that workpieces arrive at a fixed reference position is combined with its current location into a set of coordinates to uniquely identify a part. So the duplicate image information can be found and removed by comparing those coordinates periodically. At the same time, in order to improve sorting efficiency, a dynamic picking algorithm based on Newton-Raphson method is proposed. The non-linear mathematical model is established for workpiece tracking, which is solved by Newton-Raphson iteration. Finally, the proposed dynamic picking algorithm is verified by MATLAB. In prototype test the maximum sorting speed can reach 110 times per minute, mistaken-grab rate is lower than 2‰, missing-grab rate is 0, which proves that the algorithms can meet the real-time, the accuracy and the stability requirements.

     

/

返回文章
返回