基于凸包特征象棋棋子识别系统设计

摘要

Aiming at the problem of high complexity and low efficiency of character recognition process in general chess pieces vision system,a kind of chess character recognition system is designed based on convex hull features.The moving object is extracted according to background subtraction,hoff circular test is taken to determine whether there are chess pieces.Using color histogram to distinguish the chess pieces,preprocessing is performed on word morphological skeletonization,etc.The chess model recognition is realized by the characteristics of convex hull.The experimental results show that the designed system has high chessman recognition accuracy to avoid chessman recognition in terms of tilt and camera distortion,improve the recognition accuracy and efficiency,as well as provide new ideas for a text field quickly identify.%针对象棋棋子识别系统复杂度高且效率低的问题,提出了一种基于凸包特征象棋文字识别系统的设计方法.该方法根据背景减法提取运动目标,霍夫圆检测判定是否为棋子;利用颜色直方图区分棋子颜色,接着对文字形态学骨架化等预处理;根据凸包特征模型识别象棋文字.结果表明:此系统能够实现象棋棋子的识别任务,避免了棋子识别中文字倾斜和相机畸变等问题,识别正确性及效率明显提高,为特殊领域文字识别提供了新思路.