当前位置>主页 > 期刊在线 > 计算机技术 >

计算机技术21年24期

基于卡尔曼滤波和 YOLOv5 绘制滑行轨迹的方法
何金平,杨波
(长春财经学院信息工程学院,吉林 长春 130122)

摘  要:真实场景中的目标跟踪问题一直以来都是研究热点,而双人花样滑冰作为冬奥会热门比赛项目之一,如何检测跟踪并预测绘制运动员在比赛或训练中的曲线对于评判运动员的表现尤为重要,为此,文章提出一种基于卡尔曼滤波和 YOLOv5绘制双人花样滑冰滑行轨迹的方法。首先通过YOLOv5检测当前帧中的待跟踪目标,获取目标的ID、边界框大小和中心坐标信息;然后利用卡尔曼滤波根据目标的上一帧位置信息预估其下一帧位置,进行精确的目标跟踪;最后绘制出运动员的运动轨迹。


关键词:卡尔曼滤波;YOLO;目标检测;花样滑冰;轨迹



DOI:10.19850/j.cnki.2096-4706.2021.24.039


中图分类号:TP18                                      文献标识码:A                                  文章编号:2096-4706(2021)24-0153-06


A Method of Drawing Taxiing Trajectory Based on Kalman Filter and YOLOv5

HE Jinping, YANG Bo

(School of Information Engineering, Changchun University of Finance and Economics, Changchun 130122, China)

Abstract: Target tracking in the real scene has always been a research hotspot. And double figure skating taking as one of the popular events in the Winter Olympic Games, how to detect, track, predict and draw the curve of athletes in competition or training is particularly important to judge the performance of athletes. Therefore, this paper proposes a method to draw the sliding track of double figure skating based on Kalman filter and YOLOv5. Firstly, the target to be tracked in the current frame is detected by YOLOv5, and the target ID, bounding box size and center coordinate information are obtained; then, Kalman filter is used to estimate the position of the next frame according to the position information of the previous frame of the target for accurate target tracking; finally, draw the movement track of the athlete.

Keywords: Kalman filter; YOLO; target detection; figure skating; trajectory


参考文献:

[1] 孙茁鸣 . 青少年花样滑冰运动员年龄特征与机能训练的探讨 [J]. 当代体育科技, 2021,11(10),46-48.

[2] 付彩姝,王旋,姜海兰 . 花样滑冰青年运动员体能训练的方法和手段 [J]. 冰雪运动,2006(2):28-29.

[3] GUAN S F,LI X F. WITHDRAWN: Moving target tracking algorithm and trajectory generation based on Kalman filter in sports video [J/OL].Journal of Visual Communication and Image Representation,2019[2021-10-05].https://doi.org/10.1016/ j.jvcir.2019.102693.

[4] 尹宏鹏,陈波,柴毅,等 . 基于视觉的目标检测与跟踪综述 [J]. 自动化学报,2016,42(10):1466-1489.

[5] 任珈民,宫宁生,韩镇阳 . 基于 YOLOv3 与卡尔曼滤波的多目标跟踪算法 [J]. 计算机应用与软件,2020,37(5):169-176.

[6] 江金洪,鲍胜利,史文旭,等 . 基于 YOLO v3 算法改进的交通标志识别算法 [J]. 计算机应用,2020,40(8):2472-2478.

[7] 童燮,万茂,赵奉奎 . 基于目标运动轨迹预测的自动紧急制动避撞策略 [J]. 软件导刊,2021,20(4):111-116.

[8] 余弦 . 基于轨迹的视频运动对象的检测与跟踪技术研究 [D]. 上海:上海交通大学,2010.


作者简介:何金平(2000—)男,汉族,重庆忠县人,本科在读,研究方向:机器学习;杨波(1976—)男,汉族,吉林长春人, 副教授,硕士,研究方向:智能评价系统、数据挖掘。