依据SIFT算法的排球运动员急停起跳动作识别方法  

SIFT Algorithm-based Recognition Method for Volleyball Players’Emergency Stopping and Jumping Movements

作　　者：侯皖东 HOU Wandong(Department of Physical Education,Anhui Wenda University of Information Engineering,Hefei 230000,China)

机构地区：[1]安徽文达信息工程学院体育教学部,安徽合肥230000

出　　处：《新乡学院学报》2023年第12期31-35,41,共6页Journal of Xinxiang University

基　　金：2021年安徽文达信息工程学院科研基金项目重点项目(XSK2021A13)。

摘　　要：目前常规的动作识别方法主要通过提取运动图像中的动作连续帧序列,构建出动作识别模型,由于识别动作的特征描述较为复杂,识别效率较低。对此,提出基于尺度不变特征转换(SIFT)算法的排球运动员急停起跳动作识别方法。通过对图像边界点进行膨胀操作处理,获取识别对象的运动区域,并构建人体有向时空骨架图,结合SIFT算法对识别动作特征进行简化提取处理,通过对动作轮廓波域进行平滑处理,获取识别动作的轮廓曲线,实现动作识别。在实验中,对提出的方法进行了识别效率的检验。结果表明,采用提出的方法对图像进行识别处理时,动作识别延迟较低,具备较高的识别效率。The current conventional action recognition methods mainly construct action recognition models by extracting sequences of consecutive frames of actions in motion images,which leads to low recognition efficiency due to the complex feature description of recognized actions.In this regard,the SIFT algorithm-based action recognition method for volleyball players’emergency stop and jump is proposed.The motion region of the recognized object is obtained by processing the image boundary points for expansion operation.And construct the human body directed spatio-temporal skeleton map,combine the SIFT algorithm to simplify the extraction process of the recognized action features,and finally obtain the contour curve of the recognized action by smoothing the action contour wave domain to realize the action recognition.In the experiment,the recognition efficiency of the proposed method is examined.The experimental results show that when the proposed method is used for image recognition processing,the action recognition delay is low and has a high recognition efficiency.

关 键 词：急停起跳 动作识别 边缘检测 轮廓提取 SIFT算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] G841[自动化与计算机技术—计算机科学与技术]