摘要
在研究纹理特征属性的基础上,提出采用灰度共生矩阵纹理特征进行目标跟踪的方法。基于OpenCV平台,通过对标准测试视频的仿真试验,对灰度共生矩阵纹理特征、局部二进制模式纹理特征和灰度颜色直方图特征在粒子滤波目标跟踪框架中的作用进行了测试与对比分析,灰度共生矩阵纹理特征在相似性颜色的遮拦与抗扰动、处理时间等方面表现出良好的属性特征,相关的试验数据和对比结果表明了这种纹理特征具备优良的跟踪特性,可以增强跟踪系统的整体性能。
Based on the texture characteristic,the object tracking method which integrates gray level co-occurrence matrix texture＇s characteristic into system＇s framework is proposed.By doing experiments on standard video under OpenCV,the functions about the gray level co-occurrence matrix texture,local binary model texture and gray color characteristic are tested in contrastive experimentations using a particle filter tracking algorithm.The capabilities about resisting analogous color＇s disturbance and working time are embodied in experimentation using gray level co-occurrence matrix texture.The excellent quality of the gray level co-occurrence matrix texture is indicated by correlative testing data and contrastive results,which can boost up the holistic capability of the tracking system.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2010年第4期459-463,共5页
Journal of Nanjing University of Science and Technology
基金
国家自然科学基金(90820306)
关键词
目标跟踪
纹理
灰度共生矩阵
object tracking
textures
gray level co-occurrence matrix
