人机象棋对弈中棋子的快速捕捉与定位研究

发布时间：2018-01-11 23:13

本文关键词：人机象棋对弈中棋子的快速捕捉与定位研究　出处：《华中科技大学学报(自然科学版)》2017年10期 　论文类型：期刊论文

【摘要】：针对人机象棋对弈中变动棋子的快速捕捉问题,提出了一种在动态环境下快速捕捉变动棋子的方法.首先,采用背景差分法判断出棋盘上存在运动目标,使用肤色检测和人手跟踪获取人手运动轨迹,根据运动轨迹变化判断出人手已经落子.然后,获取最新的帧图像与背景图像的差分图像,帧图像再经过肤色检测处理得到对应的运动目标图像,运动目标图像与差分图像进行差分获取最终差分图像.最后,扫描最终差分图像获得两小块白色区域坐标,在提取的帧图像中相同坐标处定位识别棋子,在人手尚未离开棋盘时完成捕捉过程.采用该方法与全棋盘扫描法进行对比实验,实验结果表明:该方法可以在各种人手操控棋子的情况下可靠地捕捉到变动棋子;相比于普通的全棋盘扫描法,棋子捕捉的效率提高了36.27%.
[Abstract]:To quickly capture the problem changes in man-machine chess pieces, put forward a method in the dynamic environment changes quickly capture pieces. Firstly, we use the background subtraction method to judge the existence of moving targets on the chessboard, using skin color detection and hand tracking to obtain human hand motion track, judged according to the trajectory changes were Lazi. Then, get the new image and the background image difference image, image after color detection and processing of motion target images, moving target images and the difference image difference to obtain the final difference image. Finally, the final scan difference image obtained two small white area coordinates in the frame image extraction in the same coordinate locating piece, complete the capture process in the hand has not yet left the board. The comparative experiments were carried out with the full board scanning method, the experimental results table Ming: this method can reliably capture various chessmen in the case of various manipulations of chess pieces. Compared with the general full chessboard scanning method, the efficiency of chess capture is increased by 36.27%..

【作者单位】：武汉理工大学机电工程学院;
【基金】：国家自然科学基金资助项目(61373110,61672396) 武汉理工大学自主创新研究基金资助项目(175204001)
【分类号】：TP242;TP391.41
【正文快照】： 中国象棋机器人集合了机械视觉、图像处理、计算机控制、机械运动等多种领域的知识,一直是研究的热点之一[1].在人机对弈过程中,为了减少人的等待时间,快速捕捉到变动棋子变得至关重要.传统捕捉变动棋子的方法主要是全棋盘扫描法,在人手移动棋子完成并离开棋盘后,进行全盘扫描

【相似文献】