基于ORB算法的象棋快速识别和定位系统研究

以象棋装配为研究背景,利用ORB算法实现棋子的识别,提出根据最佳匹配特征点的像素坐标,利用几何平均坐标法求出目标棋子的坐标,利用特征点构造的向量计算目标棋子相对于模板的旋转角度,实现棋子的定位,并利用工业机器人完成象棋的装配。实验结果表明:计算出的旋转角度可以保证摆放棋子时与模板为同一方向。ORB算法的运行时间均在0.2 s以内,快于SIFT和SURF算法,可以保证系统能准确地识别和定位出目标棋子,并满足象棋装配过程中的实时性要求。

Research of fast recognition and positioning system of chess based on ORB algorithm

Taking the chess assembling as the research background, the ORB algorithm is used to recognize the chesspiece. A positioning method is proposed that the best matching feature points can be used to calculate the chesspiece coordinates by geometric mean coordinate method and rotation angle by the vector of feature points. The industrial robot is used to assemble the chess finally. The results show that rotation angle can guarantee that the chesspiece is placed in the same direction as the template. The running time of ORB algorithm is within 0.2s which is faster than SIFT and SURF. It ensures the system can recognize and position the chesspiece accurately and meet the real-time requirements of chess assembling.