摘要
本文提出了一种基于图像分割算法的篮球对象提取方法。针对篮球视频中篮球对象唯一性的特点,对视频图像进行混合高斯建模,采用Grab Cut算法对篮球对象进行分割,实现对篮球前景的提取。仿真实验结果表明对篮球视频中篮球对象分割提取具有良好的效果,且该方法具有交互操作少,分割精确度高的优点,能够推广到彩色图像领域。
An approach to the basketball extraction based on image segmentation algorithm is presented, in this paper. For the basketball object is unique in basketball game. In the proposed approach, the Gaussian Mixture Model is used to model the color data of the video frame, and the Grab Cut algorithm is adopted to segment the basketball object, in this way, the basketball foreground extraction can be achieved. The experimental results show that the effects of basketball extraction of basketball video are well by this approach, it needs less interactive operations and has high segmentation precision, and it can be applied in the field of colorful image.
出处
《微计算机信息》
2011年第4期210-212,共3页
Control & Automation
基金
基金申请人:马永伟
项目名称:篮球视频分析及检索的若干技术研究
项目代号:(B.16-0114-09-024)
项目颁发部门:上海大学研究生创新基金(SHUCX102307)
关键词
景建模
图像分割
前景提取
混合高斯模型
GRAB
Cut算法
Background Model
Image Segmentation
Foreground Extraction
Mixture Gaussian Model
Grab Cut algorithm