基于计算机视觉的艺术体操轨迹跟踪研究

发布时间：2018-06-19 14:41

本文选题：计算机视觉 + 艺术体操　；参考：《现代电子技术》2017年19期

【摘要】：为了解决传统基于卡尔曼滤波算法进行艺术体操轨迹跟踪时存在的跟踪漂移以及跟踪效率低等问题,研究基于计算机视觉的艺术体操轨迹跟踪方法,通过Vi Be运动目标检索算法对图像的颜色以及深度信息建模,基于图像颜色以及深度的波动检测出视频中的运动目标,采用KCF算法实现运动目标的初步跟踪,在该方法的基础上,通过改进KCF算法解决运动目标被遮挡出现的跟踪漂移问题,提高运动目标跟踪的精度和稳定性。通过Hermite插值运算运动目标质心,基于时刻t的运动模糊方向获取瞬时质心轨迹,得到最佳的运动目标质心轨迹,采用曲线拟合措施获取精确的运动目标质心轨迹。实验结果说明,所提方法可准确跟踪艺术体操运动轨迹,具有较高的跟踪效率和稳定性。
[Abstract]:In order to solve the problems of track drift and low tracking efficiency in traditional track tracking of rhythmic gymnastics based on Kalman filtering algorithm, the track tracking method of rhythmic gymnastics based on computer vision is studied. The color and depth information of the image is modeled by Vi be moving target retrieval algorithm. Based on the fluctuation of color and depth of the image, the moving object in the video is detected. The KCF algorithm is used to realize the initial tracking of the moving object. On the basis of this method, the tracking drift problem is solved by improved KCF algorithm, which can improve the accuracy and stability of moving target tracking. By Hermite interpolation, the instantaneous centroid trajectory is obtained based on the motion fuzzy direction of time t, and the best trajectory of moving object centroid is obtained, and the accurate trajectory of moving object centroid is obtained by curve fitting. The experimental results show that the proposed method can track the track of rhythmic gymnastics accurately and has high tracking efficiency and stability.
【作者单位】：重庆三峡学院;
【分类号】：G834;TP391.41

【相似文献】