摘要
一般的Snake模型用于目标跟踪时存在一些问题,如要求初始控制点位置非常接近于真实 轮廓、不能可靠跟踪快速运动目标、对目标的运动规律有许多约束条件等。在继承所提出的Auto-Snake模型跟踪算法优点的基础上,进一步采用最小二乘估计预测器对目标的位置进行预测,放宽对运动规律的约束,可以实现对快速运动(大位移)的目标的自动跟踪。
Some problems such as initial control point position very nearing real profile, impossible reliably tracking fast moving targets and many restraint conditions for moving rules of a target, etc existed when common Snake model is applied to target tracking. The target position is further predicated by least-squares estimation on the basis of inheriting the proposed Auto-Snake model-tracking algorithm. By relaxing the restraints for moving rules, the automatic tracking for fast movement (large displacement) can be carried out.
出处
《光电工程》
CAS
CSCD
北大核心
2003年第1期20-23,共4页
Opto-Electronic Engineering
