Research on Chinese chess automatic notation method
Addressing the issue of cumbersome and costly existing notation methods for Chinese chess,an automatic chess notation method based on machine vision is proposed.The method involved a series of steps,beginning with image preprocessing,followed by hand occlusion detection using binarization and connected region search.Subsequently,Hough circle detection,character matrix and so on were combined to locate the chess pieces.The next steps involved classifying the chess pieces into red ones and black ones,distinguishing between them and employing the local binary pattern histogram(LBPH)algorithm for chess species recognition.Finally,the moves were generated by dynamically recognising the movement paths of the chess pieces.To evaluate efficacy of the method,50 chess match videos were selected for testing.The results demonstrate that this method can process and recognise 5 frames per second with 99%accuracy,while achieving a 100%success rate for producing notations across all tested videos,which can fully meet the notation requirements for various types of games.
国家科技期刊平台
NSTL
万方数据
维普
